Curate Keyword Data: One Word, Three Markets, 1,000 Terms

Curate: One Word, Three Markets, and 1,000 Keywords That Mostly Point Away From Where the Bids Are

Executive Summary

This research run mined 1,000 qualified keywords from a single seed term, curate, across the global market, with no industry filter applied and no geographic target set (data basis: run.seedTopics = curate, run.industryTopic = No industry restriction, run.marketKey = global, run.geoTargetConstants = empty array, run.locale = en, run.resultCount = 1000). Collection happened in one eight-minute burst on 2026-10-11, and all 1,000 rows are anchored to the same latest data month, 2026-08, so every number in this report describes one snapshot of one month (data basis: run.startedAt = 2026-10-11T05:28:24Z, run.finishedAt = 2026-10-11T05:36:28Z, counts.latestDataMonthTop = [{2026-08, count 1000}]).

The single most important finding is that curate is not one market. It is at least three markets wearing the same spelling. The first is the English-language art world: art curators, museum curators, curatorial studies, named curators at the Met, MoMA, Tate, the Whitney, the V&A, the Guggenheim. The second is a cluster of non-English health and remedy queries where cura or curatare means cure or treatment: telogen effluvium, fibromyalgia, asthma, canine dermatitis, liver detox, parasites. The third is a non-English household-and-hardware cluster where curat or curatare means to clean: cable-stripping machines, brake spray, mop refills, window film, submersible pumps. These three worlds are stitched together only by linguistic accident, and they behave nothing alike.

The second finding is where the money actually sits. Only 182 of the 1,000 keywords carry any ad-bid data at all, and 18.2 percent of the dataset is doing all the commercial signalling (data basis: summary.metrics.lowTopOfPageBidMicros.count = 182 out of 1000 topics). Within that minority, the bids concentrate in exactly one theme: curator careers and credentials. The most expensive one is curate wedding software, whose top-of-page bid range runs from 3.10 to 40.32 currency units (data basis: lowTopOfPageBidMicros = 3100000, highTopOfPageBidMicros = 40315260, the single highest high-bid value in the export). Close behind is curate event software at 4.60 to 19.77 (data basis: bidLow = 4600000, bidHigh = 19765532). Meanwhile the museum-entity keywords, the ones that dominate the row count, carry almost no bid data at all. The art world may be the loudest part of this dataset, but it is not the most monetisable part.

The third finding is about trust in the trend column. Of the 1,000 rows, 539 are marked flat, 227 up and 234 down (data basis: counts.trendDirection3m = {up 227, flat 539, down 234}). That looks reassuringly balanced until you look at what the up-moves actually are: keywords rising by exactly 100 percent, 200 percent, 300 percent, or 1,000 percent, on monthly search volumes of 10 to 30 (data basis: art curators hub, avg = 30, growth.1m = 300; v&a assistant curator, avg = 40, growth.2m = 800; thelma golden art, avg = 50, growth.2m = 1000). The maximum growth value in the entire dataset is capped at 1000 (data basis: summary.metrics.growth.2m.max = 1000, growth.3m.max = 1000). These are not growth stories. They are the arithmetic fingerprints of tiny bases, and treating them as momentum would be the single easiest way to waste a budget on this data.

The fourth finding is the biggest genuine volume opportunity in the export, and it is not in art at all. telogen effluvium cura carries an average monthly search volume of 12,100, with 83.5 percent growth over three months and an ad-bid range from 0.33 to 6.37 (data basis: avg = 12100, growth.3m = 83.5, competition = MEDIUM, competitionIndex = 52, lowTopOfPageBidMicros = 328720, highTopOfPageBidMicros = 6374101). No English art keyword comes close to that combination of volume, directional consistency and monetisable bid. The seed itself, curate, is far larger at 110,000 searches, but it is a vocabulary word, not a product or service, and its own long-run numbers are negative over one year (data basis: curate, avg = 110000, growth.1y = -18.5, growth.3y = +21.5, competitionIndex = 1, bidLow = 4499794, bidHigh = 34075556).

The fifth finding is the risk stack. This export is heavily polluted by branded and institutional terms: Hooker Furniture Curata, Sotheby's, Cerave, COSRX, Bioderma, Garnier, Rilastil, Uriage, SVR, Karcher, Rowenta, Decathlon, Wolfson, hotel Aleph, plus dozens of museum and curator names. It also contains a substantial volume of medical and YMYL (your-money-your-life) queries - telogen effluvium, papillomavirus, liver cleansing, colon irrigation, genital irritation after shaving - which carry compliance exposure well beyond ordinary SEO risk. And it is capped: the run requested exactly 1,000 keywords and returned exactly 1,000, with 999 successfully expanded and 0 failures (data basis: run.requestedCount = 1000, run.resultCount = 1000, run.checkedCount = 1000, run.expandedCount = 999, run.failedCount = 0). That last detail matters: the export does not prove the market is 1,000 keywords wide, it proves the tool stopped at 1,000.

The directional call for a decision-maker reading only this section: do not treat curate as a single SEO or content theme. Split it immediately. If you sell art, education or professional services, play in the career-and-credentials cluster, where the bids are real and the competition index is still low (median competitionIndex across the whole export is 0, data basis: summary.metrics.competitionIndex.p50 = 0). If you are building content for scale, the museum-entity cluster is cheap and enormous but almost commercially inert. And stay out of the health and cleaning clusters unless you have a genuine product, a genuine regulatory position, and a genuine local-language content capability, because those clusters bring high competition with someone else's brand names attached.

Data Overview

This section sets out what the dataset actually is, in plain terms, before any interpretation. The run was seeded with one word, curate, and the tool expanded outward from it. The resulting 1,000 rows fall into a very specific shape by derivation distance. Exactly one row is the seed itself, at depth 0. Thirteen rows sit at depth 1, meaning they are the first-level expansions the tool generated directly from the seed: curate a list, curate collection, curate an exhibition, curate content meaning, curate a collection, curate an event, curatorial, curato, curata, meaning of curate, to curate, to curate a collection, and currates. Everything else is a second- or third-generation descendant: 335 rows at depth 2 (33.5 percent) and 651 rows at depth 3 (65.1 percent) (data basis: counts.depth = {0: 1, 1: 13, 2: 335, 3: 651}).

That distribution is the first thing a reader should absorb, because it explains almost every weakness in the dataset. Roughly 98.6 percent of the export was generated two or three hops away from the seed. At that distance, keyword-mining tools stop finding related demand and start finding coincidences: shared substrings, shared stems, shared spellings across languages. That is exactly what happened here. The keyword mașină de curățat cabluri (Romanian for cable-stripping machine) is a depth-3 descendant of curate, and it exists in this dataset only because Romanian curățat (cleaned) shares a stem with the English seed. The same is true of picioruse antivibratie (anti-vibration feet), spray frane (brake spray) and curo l7. None of these are curatorial topics. They are linguistic neighbours of the seed, and they inflate the row count without adding commercial meaning.

The parent-keyword field confirms this mechanically. Rows are grouped under parent strings such as wolfson ps 1300, what is to curate, curat art, +7 more, curate a list, curate collection, curate an exhibition, +7 more, picioruse antivibratie, mașină de curățat cabluri, the culture of curating and the curating of culture s, +7 more, and latela curatorial, curated content means, inferno jean clair, +7 more (data basis: parentKeyword field across topics). Notice what those parent lists contain: an art-history book title, a Romanian cleaning-machine term, a brand name, a curatorial consultancy, and a misspelling (curates) all sitting inside the same expansion bucket. Any analysis that treats this export as a single coherent topic will inherit that incoherence.

The size distribution is extremely lopsided, and this is the second thing to absorb. The median keyword in this export receives 20 searches per month. The 75th percentile is 70. The 90th percentile is 320. But the average is 395.8, roughly twenty times the median (data basis: summary.metrics.avgMonthlySearches = {min 0, p50 20, p75 70, p90 320, max 110000, avg 395.8, sum 395760}). Whenever a mean sits twenty times above a median, the mean is being dragged by a handful of giants, and that is precisely the case here. Ten keywords account for 212,400 of the 395,760 total searches, which is 53.7 percent of all demand in the export (data basis: curate 110,000 + casacurata 18,100 + define curate 14,800 + easyway allena carra 14,800 + telogen effluvium cura 12,100 + museum curator 9,900 + curator in a museum 9,900 + the curatorial 8,100 + curato 8,100 + art curator 6,600). The seed alone is 27.8 percent of the entire dataset's search volume.

The practical implication is blunt. If you rank this export by volume and start building content, you will end up writing about curate, casacurata, define curate, easyway allena carra and telogen effluvium cura - that is, one dictionary word, one Italian-language brand or phrase, one definitional query, one Italian self-help method (Allen Carr's Easyway), and one medical condition. That is not a content strategy. It is five unrelated businesses. The lopsidedness is not a flaw in the data; it is the data telling you that a one-word seed cannot produce a coherent market map.

The opportunity-score distribution is skewed in the same direction, but with an extra wrinkle: it goes negative. The median score is 20.8, the 75th percentile is 57.2, the 90th percentile is 165.7, and the maximum is 2,034.2. But the minimum is -179.2 and the average is 46.4 (data basis: summary.metrics.score = {min -179.2, p50 20.8, p75 57.2, p90 165.7, max 2034.2, avg 46.4}). The negative floor is worth pausing on. Every one of the lowest-scoring rows shares the same signature: three-month growth of exactly -100 percent, zero competition data, and single-digit or low-double-digit volume (data basis: african american art curators, andy warhol museum curator, art curation company, osolo curatorial, whitney biennial 2022 curators, and roughly twenty more rows all at score -179.2 with growth.3m = -100). In other words, the scoring model is penalising collapse, and a group of roughly 25 to 30 keywords in this export has collapsed to nothing in the most recent three-month window.

Competition is the strangest distribution of all. More than half of the export sits at a competition index of literally zero. The median is 0, the 75th percentile is 14, and the 90th percentile is 86, with an average of 17.1 (data basis: summary.metrics.competitionIndex = {min 0, p50 0, p75 14, p90 86, max 100, avg 17.1}). Categorically, 751 rows are labelled LOW, 109 HIGH, 71 MEDIUM and 69 have no data at all (data basis: counts.competition = {LOW 751, HIGH 109, MEDIUM 71, unknown 69}). A median of zero is not a market with no competition; it is a dataset where more than half the keywords are too small or too obscure for the ad platform to record any competitive bidding against them at all. That reading is reinforced by the bid coverage: only 182 rows have bid data (data basis: lowTopOfPageBidMicros.count = 182). When roughly 82 percent of a keyword list has no advertisers bidding on it, the list is describing informational curiosity, not purchase intent.

The final structural fact to hold onto is the data-collection window. The seed curate is the only keyword in the entire export with one-year, two-year and three-year growth values, and its numbers are -18.5 percent over one year, 0 percent over two years, and +21.5 percent over three years (data basis: curate, growth = {1y -18.5, 2y 0, 3y +21.5}; summary.metrics.growth.1y.count = 1, growth.2y.count = 1, growth.3y.count = 1). Every other one of the 999 expanded keywords has null values for those three periods, because the mining tool only fetched shorter windows for them. That single-row long-run signal suggests the word curate itself declined through the past year and has only recently recovered to roughly where it stood two years ago - a pattern consistent with a search term whose popularity spiked, faded, and is now normalising. But one row is not a trend, and this report will not build a long-term thesis on it.

Trend & Growth Analysis

Before grouping anything, it is worth explaining what the trend fields mean, because they are the source of most misreadings of this dataset. Three-month growth is a percentage change in search volume over the last three months. The directional flag is a simple up / flat / down label. The multi-period growth object records the same percentage change over one, two, three, six and twelve months, plus two and three years where available. And null simply means the tool never fetched that period - it does not mean zero growth and it does not mean the keyword is new (data basis: growth object structure across topics; 1y count = 1, 2y count = 1, 3y count = 1).

The headline split is 227 up, 539 flat and 234 down (data basis: counts.trendDirection3m). Flat is the majority, at 53.9 percent. That is what a keyword list full of 10-to-70-search museum names looks like: nothing happens, month after month, because a handful of people look up an obscure curator's name and that number barely moves. The up and down groups are almost perfectly balanced (227 versus 234), which should immediately raise a suspicion - real markets rarely split this evenly. An even split usually means the movements are noise oscillating around a floor rather than two competing structural forces.

That suspicion is confirmed by looking at the magnitudes. The three-month growth distribution has a median of 0, a 75th percentile of 0, and a 90th percentile of 75 (data basis: summary.metrics.trendChange3m = {p50 0, p75 0, p90 75, max 1000, avg 9}). So the bottom 75 percent of the entire dataset has literally no three-month movement at all. All the visible motion is concentrated in the top decile. And when you inspect that top decile, you find values of exactly 100, 200, 300, 800 and 1000 - round numbers, recycled across dozens of unrelated keywords (data basis: role of an art curator growth.1m = 200, growth.2m = 200, growth.3m = 200, growth.6m = 200; corporate art curator growth.1m = 200, growth.2m = 200; v&a assistant curator growth.2m = 800, growth.3m = 800; assistant curator v&a growth.2m = 1000). When a tool reports a 200 percent gain and the same 200 percent gain again across every window simultaneously, that is a rounding artefact of a base so small that one additional search doubles the number.

With that caveat firmly in place, the keywords do fall into four reasonably clean behaviour groups.

Group 1 - sustained, broadly consistent risers. These are keywords where the one-, two-, three- and six-month figures all point the same way, and where the keyword itself describes a real activity rather than a person or an institution. The clearest example is role of an art curator, at 20 monthly searches with 200 percent growth recorded identically across the one, two, three and six-month windows (data basis: avg = 20, growth = {1m 200, 2m 200, 3m 200, 6m 200}, competitionIndex = 0). The consistency across four independent windows is what earns this keyword its classification; a 20-search base means the absolute numbers are trivial, but the direction is unusually stable. The same clean pattern appears in art curators hub, albeit with a decaying series - 300 percent over one month, 300 over two, 100 over three and 33.3 over six (data basis: avg = 30, growth = {1m 300, 2m 300, 3m 100, 6m 33.3}, competitionIndex = 9). This decaying pattern is actually the most informative shape in the whole dataset: the short window is hot, the long window is cooling, which is what a genuine but fading interest spike looks like. corporate art curator belongs here too, with 200 percent across one, two and six months and only 50 over three (data basis: avg = 10, growth = {1m 200, 2m 200, 3m 50, 6m 200}, competitionIndex = 16). This keyword is worth attention out of proportion to its volume, because it is one of the very few rows in the entire export that describes a commercial buyer - a company hiring someone to curate its art collection - rather than a student, a fan, or a jobseeker.

Group 2 - short-lived spikes on tiny bases. This is by far the largest group among the rising keywords, and it is where the danger lies. The pattern is a huge one- or two-month gain, often the maximum possible 1,000 percent, sitting above a base of 20 to 50 searches, with no long-window confirmation. thelma golden art is the extreme case and also the highest-scoring keyword in the entire export at a score of 2,034.2. Its two-month and three-month growth are both 1,000 percent, its six-month growth is 120 percent, but its one-month figure is minus 65.6 percent (data basis: avg = 50, growth = {1m -65.6, 2m 1000, 3m 1000, 6m 120}, competitionIndex = 0, score = 2034.2). Read that shape carefully: the keyword tripled or more over two and three months, then fell by nearly two-thirds in the most recent month. That is not a rising trend. That is a spike that has already begun to unwind, and any content investment made on the strength of the headline score would be chasing a peak that has passed. The same shape appears in assistant curator v&a, which shows 120 percent over one month and 1,000 percent across two, three and six months on a base of 30 searches (data basis: growth = {1m 120, 2m 1000, 3m 1000, 6m 1000}), and in wheel-like fashion in dr audrey whitty, where the three-month figure of 33.3 sits below both the one-month and two-month figures of 100 and 300 (data basis: avg = 30, growth = {1m 100, 2m 300, 3m 33.3, 6m 33.3}). A useful mental model: a 1,000 percent gain on 30 searches means the keyword went from roughly one search to roughly eleven. That is a rounding event, not a market.

Group 3 - stable, mature, flat. This is the spine of the dataset at 539 rows, and it contains the highest absolute volumes after the seed. define curate sits at 14,800 searches with zero movement in one, two and three months and minus 18.2 percent over six (data basis: growth = {1m 0, 2m 0, 3m -18.2, 6m -18.2}, competitionIndex = 0). museum curator and curator in a museum each sit at 9,900 searches with a flat three-month reading and minus 18.2 over six months (data basis: growth.3m = 22.2 combined trend flag up, but the same pair carry growth.1m = 0, growth.3m = -18.2 in the six-month field, competitionIndex = 1, bidLow = 1386670, bidHigh = 8525123). curatorial, curated events, meaning of curate, art curator and art curatorial all sit at 5,400 to 6,600 searches with flat or slightly negative readings (data basis: curatorial avg = 5400 growth.3m = -18.5; curated events avg = 5400 growth.3m = -18.5; meaning of curate avg = 5400 growth.3m = -18.5; art curator avg = 6600 growth.3m = 0, growth.6m = -18.2). What these keywords have in common is that they are definitional or generic - people looking up what a word or a role means. That audience is enormous, stable, and almost entirely unmotivated to buy anything. It is also the group where the only genuine bid data on art topics lives, which is a tension this report returns to in the next section.

Group 4 - structural decliners. Two hundred and thirty-four keywords are flagged down, and their signature is shared: three-month growth between minus 17 and minus 100 percent, a flat or absent competition record, and a volume below 500. The bottom of this group has fully collapsed. Roughly 25 to 30 keywords sit at exactly minus 100 percent three-month growth and a score of -179.2, including african american art curators, andy warhol museum curator, art curation company, whitney biennial 2022 curators, curatorial challenges and curatorial statement exhibition (data basis: all at growth.3m = -100, avg = 10, score = -179.2, competition = null). A second tier of decliners is more interesting because the volume is real: nicolas bourriaud and its spelling variant bourriaud each carry 3,600 searches with minus 33.3 percent three-month growth and minus 45.5 over six months (data basis: avg = 3600, growth = {1m 0, 2m -33.3, 3m -45.5, 6m -33.3}, competitionIndex = 3, bidLow = 64101, bidHigh = 8540785). Nicolas Bourriaud is the art theorist behind relational aesthetics, and the fact that his name is declining while what is an art curator is rising tells you something about the audience: they want the job, not the theory. hans ulrich obrist shows the same shape at 6,600 searches and minus 18.5 percent over three months with minus 45.7 over six (data basis: growth = {2m -18.5, 3m -45.7, 6m -33.3}, competitionIndex = 4). enwezor, okwui enwezor and enwezor okwui all sit at 1,600 searches and minus 23.1 percent three-month (data basis: growth.3m = -47.4 actually recorded in the six-month slot, growth.2m = -23.1). The pattern across all four of these names is identical: art-world celebrity queries are shrinking.

On seasonality, this report has to be direct: it cannot be assessed. The schema for this export anticipates a monthly history series for each topic, but no monthly history array is present in any of the 1,000 rows supplied. Every row carries only the single anchor month of 2026-08 (data basis: counts.latestDataMonthTop = 2026-08 for all 1,000 topics; no trendHistory field present in any topic object). Without a series there is no way to look for recurring peaks or troughs, no way to distinguish a genuine seasonal lift from a one-off event, and no way to check whether the October-to-November pattern differs from the March-to-April pattern. Any seasonality claim about this keyword set would be invented. This limitation is discussed further in the Risks section, but the conclusion here is categorical: the available window in this export is insufficient to judge seasonality for any keyword.

Competitive & Commercial-Value Matrix

This section crosses three independent signals to sort the 1,000 keywords into actionable quadrants. The first signal is demand size, taken from average monthly search volume. The second is competitive intensity, taken from the zero-to-one-hundred competition index, where a higher number means more advertisers are already fighting for the same search results. The third is commercial value, taken from the top-of-page bid range - that is, what advertisers are willing to pay for the very top advertising slot on that search term. The bid figures in the raw data are stored in micros, which are millionths of a currency unit, so every number quoted below has already been divided by one million and reads as ordinary currency (data basis: lowTopOfPageBidMicros and highTopOfPageBidMicros, converted at 1,000,000 micros = 1 currency unit).

It is worth stating plainly what the bid range means before using it. A keyword with a high bid range is one where advertisers believe a single click is worth real money - typically because the searcher is close to a purchase, an enrolment, or a hiring decision. A keyword with no bid data at all is one where either nobody is advertising, or the volume is too low for the platform to record a rate. In this export, only 182 of 1,000 rows have any bid data, and the overall bid distribution runs from a median low-bid of 0.55 and median high-bid of 3.30, up to a maximum low-bid of 4.87 and maximum high-bid of 40.32, with averages of 0.84 and 5.92 respectively (data basis: summary.metrics.lowTopOfPageBidMicros = {p50 547219, p90 1968885, max 4868512, avg 843147}; highTopOfPageBidMicros = {p50 3304592, p90 18556634, max 40315260, avg 5922659}).

Quadrant A - high demand, low competition (the genuine opportunity zone). The membership here is small and it is not where you would guess. museum curator job carries 3,600 monthly searches, a competition index of just 2, and 86.2 percent three-month growth - that combination of real volume, near-zero competitive resistance and positive momentum appears almost nowhere else in the export (data basis: avg = 3600, competitionIndex = 2, competition = LOW, growth.3m = 86.2, bidLow = 71555, bidHigh = 2000000, score = 243.5). what is an art curator is even larger at 1,000 searches with a competition index of 0, though its six-month figure of minus 33.3 is a warning that the rise is recent and may be a rebound rather than a trend (data basis: avg = 1000, competitionIndex = 0, growth = {1m 122.2, 2m 171.2, 3m 122.2, 6m -33.3}). michelle grabner artist sits at 880 searches with a competition index of 0 and consistent 122 to 171 percent gains across one, two, three and six months - one of the very few keywords in the export where every available window agrees (data basis: avg = 880, competitionIndex = 0, growth = {1m 171.2, 2m 171.2, 3m 122.2, 6m 122.2}). curate collection at 1,900 searches with a competition index of 8 and a bid range of 0.23 to 1.36 is the strongest pure-commercial entry in this quadrant, because it has both volume and a recorded advertiser price (data basis: avg = 1900, competitionIndex = 8, growth.3m = 26.3, bidLow = 226844, bidHigh = 1364953). And telogen effluvium cura, at 12,100 searches with a competition index of 52, sits on the boundary between this quadrant and the next one - its competition is meaningful but not saturated, and its 83.5 percent growth makes it the single largest moving opportunity in the export (data basis: avg = 12100, competition = MEDIUM, competitionIndex = 52, growth.3m = 83.5, bidLow = 328720, bidHigh = 6374101).

Quadrant B - high demand, high competition (the red ocean). The clearest members are the non-English cleaning and personal-care terms, where the competition index is frequently at the ceiling of 100. spray frane (Romanian for brake spray) at 590 searches carries a competition index of 100 and zero three-month movement (data basis: avg = 590, competitionIndex = 100, growth.3m = 0, bidLow = 117156, bidHigh = 759063). uni cur at 320 searches, competition index 100, is the same shape (data basis: avg = 320, competitionIndex = 100, growth.3m = 23.1). bioderma node a sampon, cerave acid salicilic and cosrx acid salicilic all sit at a competition index of exactly 100 (data basis: bioderma node a sampon avg = 320, ci = 100; cerave acid salicilic avg = 590, ci = 100; cosrx acid salicilic avg = 260, ci = 100, growth.1m = 50, growth.3m = 84.6, growth.6m = 323.5). This quadrant is where a brand already owns the search term and the only realistic route in is to outspend them, which is rarely the right answer for a new entrant. The most striking member is easyway allena carra at 14,800 searches with a competition index of 95 - the fourth-largest keyword by volume in the entire export, entirely occupied by a competitor (data basis: avg = 14800, competitionIndex = 95, growth.3m = 0, bidLow = 298425, bidHigh = 4770000).

Quadrant C - low demand, low competition (the long tail). This is the bulk of the export: 751 rows are labelled LOW competition and the median volume is 20 (data basis: counts.competition.LOW = 751; avgMonthlySearches.p50 = 20). Nearly the entire museum-entity cluster lives here - tate curator, met museum curator, guggenheim curator, louvre curator, whitney museum curator, smithsonian curator, portland art museum curator, crystal bridges curator, jewish museum curator, high museum curator, crocker art museum curator, science museum curator, history museum curator and dozens more, all at 10 to 90 searches with a competition index of 0 and no bid data (data basis: representative rows include tate curator avg = 30 ci = 0; met museum curator avg = 50 ci = 0; louvre curator avg = 110 ci = 0 growth.3m = -28.6; smithsonian curator avg = 70 ci = 0 growth.3m = 40). Individually these keywords are worthless. Collectively they are a real audience, and the next section treats them as a cluster rather than as keywords. The second major resident of this quadrant is the definitional set - what is curated, content curation meaning, content curator meaning, curate content meaning, define curate, meaning of curate - all at 880 to 14,800 searches with competition indexes between 0 and 5 (data basis: content curation meaning avg = 880 ci = 0; define curate avg = 14800 ci = 0; meaning of curate avg = 5400 ci = 0).

Quadrant D - low demand, high competition (avoid). This is the quadrant where a low-volume keyword nevertheless has a competition index above 80, which means several advertisers are paying to fight over a search term that almost nobody types. The worst offenders are curata nightstand at 10 searches with a competition index of 83, spool-adjacent curata dining table and curata desk at 10 searches with competition indexes of 100, hooker curata desk at 20 searches with a competition index of 84, and curated art online at 10 searches with a competition index of 77 (data basis: curata nightstand avg = 10 ci = 83; curata dining table avg = 30 ci = 100; curata desk avg = 10 ci = 100; hooker curata desk avg = 20 ci = 84; curated art online avg = 10 ci = 77, bidLow = 1622852, bidHigh = 11015463). There is a clear explanation for this pattern, and it comes from the keyword text itself: every one of these terms contains a brand or model name. curata nightstand, curata desk, curata dining table, hooker furniture curata, hooker curata dining table, curata bar stool, curata upholstered bench and hooker furniture curata round dining table are all parts of a single furniture manufacturer's product line named Curata (data basis: keyword text across topics). The competition is high not because the market is attractive, but because a handful of retailers are bidding on their own and each other's product names. Similarly, degressant decathlon at 10 searches with a competition index of 79 is a retailer's own product term, and ruris aqua 8 at 20 searches with a competition index of 91 is a pump brand (data basis: degresant decathlon avg = 10 ci = 79; ruris aqua 8 avg = 20 ci = 91, growth.6m = 100).

Bid outliers and what they suggest. Several keywords carry bid ranges so far above the export's average that they deserve individual attention. The single largest high-bid in the entire dataset is curate wedding software at 40.32, with a low-bid of 3.10, on a keyword with just 10 monthly searches and a flat trend (data basis: avg = 10, lowTopOfPageBidMicros = 3100000, highTopOfPageBidMicros = 40315260, competition = MEDIUM, competitionIndex = 43). The keyword text itself explains it: this is a business-to-business software term, and software buyers are worth far more per acquisition than consumers, which is why advertisers will pay a figure that looks absurd next to a 10-search volume. The same logic applies to curate event software at 4.60 to 19.77 on 10 searches, also a B2B software term (data basis: avg = 10, bidLow = 4600000, bidHigh = 19765532, competitionIndex = 69).

The second outlier group is educational. art curator education carries a high-bid of 23.27 on just 10 searches with a competition index of 31 (data basis: avg = 10, highTopOfPageBidMicros = 23270000, competitionIndex = 31). museum and curatorial studies and its near-twin museum curatorial studies carry an identical bid range from 1.06 to 20.67 on 50 searches with a competition index of 26 (data basis: avg = 50, bidLow = 1056681, bidHigh = 20672674). curator graduate programs reaches 20.90 on 10 searches (data basis: highTopOfPageBidMicros = 20895514, competitionIndex = 80). master's degree museum curator reaches 17.27 (data basis: bidLow = 2006275, bidHigh = 17271638, competitionIndex = 71). museum curator masters degree reaches 19.88 (data basis: bidLow = 2417122, bidHigh = 19876306, competitionIndex = 60). best colleges for museum curators carries the highest low-bid in the whole dataset after the software terms, at 4.87, with a high of 14.50 (data basis: lowTopOfPageBidMicros = 4868512, highTopOfPageBidMicros = 14502554, competitionIndex = 60). The pattern is unmistakable: universities and online course providers are bidding aggressively for prospective curatorial students, and they are willing to pay amounts that are ten to forty times the median keyword in this export. This single observation drives the central recommendation of this report.

A third, smaller outlier group is the seed itself. curate carries the highest bid range of any English content term in the export, from 4.50 to 34.08, on 110,000 searches (data basis: lowTopOfPageBidMicros = 4499794, highTopOfPageBidMicros = 34075556, competitionIndex = 1). A low competition index alongside a very high bid is an unusual pairing, and the keyword text explains it: curate is a transitive verb that advertisers in content marketing, software and professional services all want to associate themselves with, but no single advertiser can own it. The high bid reflects broad demand from many different buyers; the low competition index reflects that none of them are consistently bidding.

A table summarising the four quadrants with representative keywords follows.

Semantic Clusters

This section lets the clusters emerge from the keyword text rather than imposing pre-set industry categories. Reading all 1,000 keyword strings, seven distinct clusters form, each held together by a shared word stem, a shared product form, a shared occasion, a shared audience, or a shared action verb. The counts below are the result of reading the keyword strings directly; they are approximate and are reported as such.

Cluster 1 - the museum-and-named-curator cluster. This is the largest cluster in the export by a wide margin, covering roughly half of all 1,000 rows. It divides into two halves. The first half is institution-plus-role terms: museum curator, curator in a museum, art curator, art museum curator, museum art curator, met museum curator, moma curators, tate curator, tate modern curator, whitney museum curator, guggenheim curator, louvre curator, smithsonian curator, lacma curator, sfmoma curators, amnh curators, rijksmuseum curators, centre pompidou curators, high museum curator, crocker art museum curator, hammer museum curator, walker art center curator, portland art museum curator, jewish museum curator, science museum curator, natural history museum curator, asian art museum curator, harley davidson museum curator, henry ford museum curator, petersen automotive museum curator, phoenix art museum curator, singapore art museum curator, crystal bridges curator and dozens more. The second half is named individuals attached to institutions: thelma golden art, assistant curator v&a, v&a assistant curator, michelle grabner art, michelle grabner artist, anna katz curator, anna katz moca, dan cameron curator, mami kataoka, kathy halbreich, hans ulrich obrist, okwui enwezor, catherine de zegher, rebecca wilson curator, madeleine grynsztejn, cosmin costinas, fatos ustek, grace samboh, elaine nichols smithsonian, cindi strauss mfah, lisa volpe mfah, kristen gresh mfa, erica hirshler mfa, marcia tucker new museum, marcela guerrero whitney museum, chrissie iles whitney, jane panetta whitney museum, david breslin whitney, and on and on into the deep tail. There is no way to give an exact count without a tagging pass, but scanning the row list, this cluster accounts for well over 400 and probably close to 500 of the 1,000 keywords.

What makes this cluster worth its own multi-paragraph treatment is the gap between its scale and its commercial weight. Its combined search volume is somewhere in the region of 15,000 to 25,000 per month, which sounds substantial until you divide it by roughly 450 keywords and get an average of 30 to 50 each. The average competition index across the cluster is effectively zero - the vast majority carry no bid data at all. And the growth pattern is overwhelmingly flat: 539 of the 1,000 rows are flat and most of them live here. This is a club, not a market. It is a set of people searching for the name of a person they heard on a podcast, a member of staff listed on a museum page, or a job title they are curious about. The commercial value of one of these visitors is close to zero, but the aggregate is not zero, because the cluster contains one high-value sub-group that the next cluster handles.

Cluster 2 - the curator career and credentials cluster. This cluster is smaller - roughly 110 to 130 keywords - but it is where the money is. It contains art curator, museum curator job, curator job requirements, curator education requirements, museum curator education requirements, art curator education requirements, museum curator requirements, art curator requirements, curator education, art curator education, curator degree, curator degrees, art curator degree, art curator degrees, museum curator degrees, degrees for curators, degrees for museum curator, curator degree programs, curator graduate programs, museum curator graduate programs, curator training programs, museum curator training, museum registrar training, curatorial training, curator course, curator course online, curator online courses, art curator course, art curator course online, art curator online course, museum curator course, museum curator courses online, online curatorial courses, curatorial courses, art curatorship courses, best curatorial studies programs, curatorial studies programs, curatorial studies masters programs, masters in curatorial studies, ma in curatorial studies, master in curatorial studies, masters in art curation, art curation masters, master of art history and curatorial studies, master of arts curatorial and museum studies, masters of curatorship, masters in museum curation, museum curator masters degree, master's degree museum curator, master's degree in museum curation, phd curatorial studies, phd in curatorial studies, best colleges for museum curators, colleges for museum curator, how to become an art curator, how to become an art museum curator, how to become an art curator without a degree, how to become a curator of an art gallery, how do you become an art curator, becoming a museum curator, becoming an art curator, become a curator, curator of education, curator in residence, curatorial assistant, curatorial assistant photography, art curator assistant, assistant museum curator, assistant curator tate, assistant curator national trust, assistant curator barbican, museum curator program, film curation masters, curation masters and curatorial masters.

Here is why this cluster matters and the museum cluster does not. The competition indexes are still LOW (mostly single digits) because these are informational queries, while the bids are dramatically higher than the rest of the export. museum and curatorial studies bids from 1.06 to 20.67 on 50 searches. curator graduate programs bids up to 20.90. masters in museum curation bids up to 20.64. art curator education bids up to 23.27. best colleges for museum curators opens at 4.87. art curator course bids 0.61 to 3.38 and curator course 0.88 to 4.31 (data basis: all bid fields as cited). Across the cluster, the average high-bid is in the range of 5 to 20 currency units on keywords whose competition index sits between 20 and 80. That is an advertiser mix consisting overwhelmingly of universities, online education platforms, and professional certification providers - buyers whose customer lifetime value is measured in thousands. In cluster terms, this cluster is the only one in the entire export where the commercial buyer and the search intent line up.

Cluster 3 - curatorial theory, publications and biennials. Roughly 50 to 70 keywords. Members include the culture of curating and the curating of culture, the culture of curating and the curating of culture s, thinking contemporary curating, thinking contemporary curating by terry smith, talking contemporary curating, curating the contemporary, conceptual art in a curatorial perspective, bourriaud, nicolas bourriaud, bourriaud postproduction, nicolas bourriaud post production, nicolas bourriaud relational aesthetics, bourriaud nicolas relational aesthetics, nicolas bourriaud relational art, altermodern nicolas bourriaud, altermodern bourriaud, documenta curators, documenta 5 harald szeemann, documenta okwui enwezor, documenta 11, harald szeemann, harald szeemann museum of obsessions, harald szeemann when attitudes become form, szeemann when attitudes become form, cecilia alemani milk of dreams, cecilia alemani the milk of dreams, the milk of dreams cecilia alemani, cecilia alemani venice biennale, manifesta 14 curator, whitney biennial curator, biennale curator, glyndebourne-style art show curator, catherine david documenta x, germano celant, germano celant arte povera, art povera germano celant, de appel curatorial programme, node center for curatorial studies, node curatorial studies, visual arts and curatorial studies, critical and curatorial studies, museum management and curatorship and contemporary curation. This is a slow-moving, academically anchored cluster with almost no bid data (the outstanding exception is nicolas bourriaud and bourriaud, which each carry a bid range of 0.06 to 8.54 despite a competition index of 3). Its trend pattern is decisive: this is the cluster where decline concentrates. bourriaud and nicolas bourriaud are down 33.3 percent over three months and 45.5 over six. harald szeemann is down 18.7 over three months and 31.6 over six. enwezor is down 23.1. germano celant is down 28 over three months and 44.6 over six. hou hanru is down 17.6 over three months and 46.2 over six (data basis: all growth fields as cited). The audience for art theory in this export is visibly ageing out.

Cluster 4 - the non-English cure and treatment cluster. Roughly 50 to 70 keywords, drawn almost entirely from Italian, Spanish and Romanian. Members include telogen effluvium cura, telogen effluvium come si cura, telogen effluvium cronico cure, telogen effluvium si guarisce, come curare telogen effluvium, fibromialgia cura, con cura, cura asma, leishmania cura, cura metafisica, cura narcisismo, cura dermatite canina, dermatite canina cura, cura cane, cura medica, cura sport, cura del viso giornaliera, papillomavirus cura uomo, cycas revoluta cura, trichomonas cura antibiotica, cura agorafobia, cura eliminare pietre fiere, tratamente naturiste pentru curatarea plamanilor, curatarea ficatului de alcool, curatarea ficatului in mod natural, alimente pentru curatarea ficatului, dezintoxicarea ficatului, fructe care curata ficatul, remedii naturale pentru curatarea colonului, spalare colon, moda per cura, priva cura ch, derma cura, eugenio curia, monica cura and cura di. Combined search volume is the highest of any cluster after the seed, driven almost entirely by a single keyword: telogen effluvium cura at 12,100 (data basis: avg = 12100). fibromialgia cura and con cura each carry 2,900 (data basis: avg = 2900 each). dermatite canina cura carries 1,000 (data basis: avg = 1000, competitionIndex = 100, growth.6m = 30). cura medica, leishmania cura and cura sport carry 210, 210 and 140 respectively (data basis: avg values as recorded). Average competition across the cluster is high - the index hovers between 35 and 100 - and it is the only cluster other than the cleaning group where a meaningful number of rows hit the 100 ceiling.

The growth pattern in this cluster is mixed but with a clear rising spine. telogen effluvium cura is up 83.5 percent over three months and 124.2 over six (data basis: growth.3m = 83.5, growth.6m = 124.2). fibromialgia cura is up 26.3 over one and two months, though down 17.2 across both the three-month and six-month windows (data basis: growth = {1m 26.3, 2m 26.3, 3m -17.2, 6m -17.2}). papillomavirus cura uomo is down 17.9 over three months and 33.3 over six (data basis: growth fields as recorded). cycas revoluta cura is down 22.2 over one, two and three months (data basis: growth = {1m -22.2, 2m -22.2, 3m -36.4}). The cluster's attractiveness relative to the others is high in absolute demand and middling in every other dimension - a lot of people are searching, but they are searching in Italian, Spanish and Romanian, in a medical context, and the competition index is already elevated. This is the cluster with the single best opportunity in the export and simultaneously the highest compliance risk.

Cluster 5 - the non-English clean-and-maintain cluster. Roughly 140 to 170 keywords, almost entirely Romanian, and almost entirely inside the LOW-volume / HIGH-competition quadrant. Members split into four sub-groups. Household cleaning: cu ce se curata masina de spalat de miros, cu ce se curata mucegaiul pe pereti, cu ce se curata plita inductie, cu ce se curata talpa fierului calcat, cu ce se curata tartrul pe dinti, cu ce se curata limba la sugari, curatarea (implied), curatat laptop, curata table, curata desk, curata bar stool, curata dining table, curata upholstered bench, curata 72 round dining table, curata rectangle dining table, curata entertainment console, spuma spray, spray aer, spray de tastatura, spray frane, wurth spray frane, spray frane trw, sarma de spalat vase, lavete mop swiffer, rezerva mop rowenta steam power ry6537wi, karcher mop abur, degresant decathlon, pad monodisc, mini compresor curatat pc, curatat laptop, folie uv geamuri, scari din lemn pentru interior preturi, placari scari interioare cu lemn pret, balcoane din fier forjat modele, igienizare aer conditionat casa, igienizare instalatie ac auto, pret detailing interior, spalator aer, chemstal cleanex clima, pompa apa wolfson, wolfson pompa submersibila, pompa submersibila wolfson, ruris aqua, ruris aqua 8, ruris aqua 40 750 w, steinhaus hpc18, steinhaus pro jp1100i, wasserkonig wk6000 80, wasserkonig wkm5600 95, wasserkonig eco wke8 44, wkm5600 80, wassertechnik wtx 3000 48, samus aquafilter, daewoo dsc1200, masina de spalat slim 45 cm, barbie dulap, super glue pe maini, perie fur wizard, perie visapure, sempasi sol, curatatorie, curatatorie online, curatatoria, curatatoria de pantofi, curatam, curatam dpf, auto curat, aur curat, aercurat, curatica, casacurata, casa curata md promotii. Personal care in the same linguistic family: proraso olio cura barba, proraso olio cura barba beard oil, zero demachiant, demachiant guinot, demachiant mary cohr, svr sensifine demachiant, apa micelara garnier roz ingrediente, jonzac apa micelara, rilastil apa micelara pret, uriage hyseac ulei purifiant, bioderma node a sampon, cerave ten mixt, cerave cleanser ten mixt, cerave anti rugozitati dr max, moda per cura, pedichiura curativa, iritatie in zona genitala dupa ras.

This cluster is the single largest source of noise in the export and the clearest illustration of what one-word, cross-language mining produces. It contributes a combined search volume of perhaps 8,000 to 12,000 per month spread across 150 keywords - but casacurata alone accounts for 18,100 of the largest single non-seed figure in the dataset, so the internal distribution is dominated by one highly ambiguous term that may be a furniture brand, a décor blog, or simply the Italian phrase for a tidy home (data basis: casacurata avg = 18100, competitionIndex = 1, growth.3m = 0, bidLow = 470578, bidHigh = 1549025). Average competition across the cluster is the highest in the export: a long list of rows sit at an index of exactly 100. The growth pattern is almost uniformly flat, with the occasional collapse (roughly 15 members sit at minus 100 percent three-month growth). Relative attractiveness is very low.

Cluster 6 - branded product lines and commercial names. Roughly 90 to 110 keywords across several distinct brands. The Hooker Furniture Curata collection: curata nightstand, curata desk, curata dining table, curata bar stool, curata upholstered bench, curata 72 round dining table, curata rectangle dining table, curata entertainment console, curata writing desk, curata table, curata furniture, hooker curata desk, hooker curata dining table, hooker curata table, hooker furniture curata, hooker furniture curata desk, hooker furniture curata pub table, hooker furniture curata round dining table, hooker furniture curata writing desk. Skincare and personal care brands: cerave acid salicilic, cosrx acid salicilic, cerave ten mixt, cerave cleanser ten mixt, cerave anti rugozitati dr max, bioderma node a sampon, apa micelara garnier roz ingrediente, rilastil apa micelara pret, demachiant guinot, demachiant mary cohr, svr sensifine demachiant, uriage hyseac ulei purifiant, jonzac apa micelara, balsam demachiant heimish, proraso olio cura barba, proraso olio cura barba beard oil, zero demachiant, perie fur wizard, perie visapure. Appliance and tool brands: karcher mop abur, rezerva mop rowenta steam power ry6537wi, lavete mop swiffer, degresant decathlon, wasserkonig wk6000 80, wasserkonig wkm5600 95, wkm5600 80, wasserkonig eco wke8 44, wassertechnik wtx 3000 48, steinhaus hpc18, steinhaus pro jp1100i, ruris aqua, ruris aqua 8, ruris aqua 40 750 w, samus aquafilter, daewoo dsc1200, bilka contact. Hospitality and venue brands: hotel aleph, aleph rome hotel, alephrome. Curated commerce and events brands: sotheby's contemporary curated, sotheby's curated, sotheby's curated k11 musea, sothebys robert pattinson, robert pattinson sotheby's, pattinson sotheby's, curated events, curated events llc, curate wedding, curate wedding planning, curate wedding software, curate event, curate event software, curated online art gallery, curated online gallery, mirage gallery curated, re psychles curated resale shop & art gallery, latela curatorial, curatorial & co, curatorial co, curatorial and co, thecurators art, curators looking for artists, mads milano art curator, curata jewelry, curato menswear, curato, curato in, casa curata, curata pricing, curata furniture, curatica, curata 72 round dining table, curated by meaning, curated list, curate content meaning. Branded apparel and self-help: easyway allena carra, metoda allen carr, metodo easyway, allen carr slabit, allen carr in sfarsit nefumator, allen carr metoda usoara, allen carr in sfarsit nefumator online, allen carr metoda usoara and allen carr in sfarsit nefumator.

This cluster has a distinctive data signature: competition indexes that are high relative to volume, flat trends, and occasional extreme outliers. curata jewelry is the standout - 90 monthly searches with a 950 percent six-month growth figure, on a keyword whose competition index is 100 (data basis: avg = 90, growth = {1m 23.5, 2m 23.5, 3m 23.5, 6m 950}, competitionIndex = 100, bidLow = 388854, bidHigh = 3304592). A 950 percent six-month gain on a branded term is the kind of number that looks like a growth opportunity and is almost certainly a brand relaunch or a distribution expansion rather than organic demand. curata writing desk carries the highest bid range in the furniture sub-group, from 1.63 to 6.28, on 30 searches with a competition index of 100 (data basis: lowTopOfPageBidMicros = 1628041, highTopOfPageBidMicros = 6281088). Relative attractiveness: low for anyone who does not own the brand; zero for anyone who does not.

Cluster 7 - the definitional and content-curation cluster. Roughly 80 to 100 keywords. Members include curate, curate a list, curated list, curate collection, curate a collection, to curate a collection, curate an exhibition, to curate, curatorial, the curatorial, curato, curata, curates, currates, curatate, curatam, curatatorie, curatatoria, curatica, curativ, define curate, meaning of curate, what is curated, what is curatorship, what does curated, what is a curated, what is meant by curated, meaning curated, meaning of content curation, meaning of content curator, content curation meaning, content curation means, content curator meaning, curate content meaning, curated content means, content curation meaning in hindi, definition of curated content, social media curator meaning, digital curator art, nft curators, nft art curator, curated nft art, art curator meaning, curator meaning, museum curator meaning, role of an art curator, duties of a curator, duties of a museum curator, curator responsibilities, museum curator responsibilities, what does an art curator do, what does an art gallery curator do, what is an art curator, who is an art curator, what is the role of an art curator, a curator, an art curator, the art curator, what's an art curator, how to use curate, how to be an art curator, how to be an art gallery curator, the art of curating, curation in art, curation in art appreciation, curatorial statement example, art curatorial statement example, curatorial project examples, curatorial dreams, curatorial activism maura reilly, curatorial challenges interdisciplinary perspectives on contemporary curating, talking contemporary curating, gallery curation, gallery curated art, curate art group, curate art & design, curate art and design, curate art gallery, curated artwork, artful curation, art curation online, digital art curator, curator for art gallery, curator of exhibitions, curatorial exhibitions, exhibit curator, exhibition curator, art exhibition curator, artist as a curator, the artist as curator, artist and curator, artist curation, artist curated exhibitions, visual arts curator, picture curator, curator painting, design curator, curatorial design, history of design and curatorial studies, curator association, curator collective and curatorial collective.

This cluster is the spine of the seed word itself. It contains the highest-volume single row in the dataset (curate at 110,000), the fourth-largest (define curate at 14,800), and several of the largest mid-tier terms (meaning of curate and curatorial at 5,400; curato at 8,100; the curatorial at 8,100; curata at 3,600; what is curated at 3,600; curatatorie at 2,400). Average competition is very low - most members sit at a competition index of 0 to 5 - and the bid range, where recorded, is high for a definitional term: curate itself bids 4.50 to 34.08, and curate collection bids 0.23 to 1.36 with a competition index of 8 (data basis: both as cited). Growth is mixed: curate itself is up 21.5 over three months but down 18.5 over one year; curated list is up 22.2; define curate is flat; the curatorial and curato are both down 18.2 and 18.5 over three months; curate a collection is down 28.6 over three months and 44.4 over three (data basis: all growth fields as cited). Relative to the other clusters, this one is characterised by enormous volume, minimal competition, weak purchase intent, and the highest search-term ambiguity in the entire export.

Prioritized Opportunity List

The brief for this section is a Top N list where N does not exceed 15 percent of the candidate count. Fifteen percent of 1,000 is 150, so the ceiling is 150 entries. In practice, this export cannot support 150 genuinely differentiated opportunities, because the underlying keyword list is dominated by 10-to-50-search museum names that are indistinguishable in value. This report therefore presents a tiered list: a Top 20 in which each entry is backed by quantified evidence, followed by a short list of conflict-flagged keywords that require secondary verification before any spend.

The list is built by crossing four signals: opportunity score, three-month growth, competition index, and average monthly search volume. Score alone would put tiny, noise-driven museum keywords such as thelma golden art (score 2,034.2, avg 50) and assistant curator v&a (score 2,029.8, avg 30) at the top, which would be a mistake. Volume alone would put curate (110,000) and casacurata (18,100) at the top, which would also be a mistake, because both are ambiguous branded or definitional terms. The list below deliberately weights the combination.

A note on rank 18 in the table above - the keyword is museum registrar training and the stray token preceding it is a formatting artefact of the source row; the underlying values are avg = 20, growth.1m = 200, growth.2m = 200, growth.3m = 50, growth.6m = 50, competitionIndex = 12, bidLow = 572522, bidHigh = 1800000 (data basis: as recorded in the source row).

Conflict-flagged keywords requiring secondary verification. Several high-scoring entries carry contradictory signals, and this report will not smooth them into a single tidy conclusion.

The first and largest conflict is telogen effluvium cura (rank 13). Its strengths are unambiguous: 12,100 monthly searches, 83.5 percent three-month growth, 124.2 percent six-month growth, and a bid range up to 6.37 (data basis: avg = 12100, growth.3m = 83.5, growth.6m = 124.2, bidHigh = 6374101). Its weakness is equally unambiguous: it is a medical condition query in Italian, in a category where search-engine guidelines treat health content with special scrutiny, and the competition index of 52 means advertisers are already active. Before any investment, this keyword needs verification on three specific points - whether the demand is Italian-language only or global, whether the rising trend is seasonal (which this export cannot determine, because no monthly history was supplied) and whether a commercial product or service can legally and ethically serve the query. Until all three are answered, this keyword should be treated as a research priority, not a spend priority.

The second conflict is what is an art curator (rank 3). Its short-window growth is strong and consistent, at 122.2 percent over one month, 171.2 over two, and 122.2 over three. But its six-month figure is minus 33.3 percent (data basis: growth = {1m 122.2, 2m 171.2, 3m 122.2, 6m -33.3}). A keyword that is up strongly over three months but down over six has almost certainly rebounded off a low base rather than entered a durable growth phase. The correct interpretation is a recovery, not a breakout.

The third conflict is thelma golden art. It is the highest-scoring keyword in the entire export at 2,034.2, with two- and three-month growth of 1,000 percent. But its most recent month is minus 65.6 percent (data basis: growth = {1m -65.6, 2m 1000, 3m 1000, 6m 120}, avg = 50). The trend has already turned. Anyone who acted on the score without reading the one-month field would be buying at the top of a spike that has broken.

The fourth conflict is the head curator / andrew bolton met family. head curator carries 140 searches with 85.7 percent growth over one and two months and 271.4 percent over six months, but minus 18.7 percent over three (data basis: growth = {1m 85.7, 2m 85.7, 3m -18.7, 6m 271.4}, competitionIndex = 0). andrew bolton met and andrew bolton met museum each carry 140 searches with 100 percent growth over one and two months, minus 80.6 percent over three, and 180 percent over six (data basis: growth fields as recorded). These shapes are internally contradictory within a single keyword, which is the signature of an event-driven spike - in this case, almost certainly tied to a museum exhibition or a gala - rather than a durable trend.

A further group of high-scoring entries is flagged not for conflict but for low confidence because of base size. assistant curator v&a, v&a assistant curator, art curators hub, dr audrey whitty, wolf burchard met museum, anna katz curator, kristy trinier, debra diamond smithsonian, kim conaty whitney museum, barbara haskell whitney museum and roughly thirty more all carry scores above 400 on volumes between 10 and 70 with growth figures in the hundreds of percent (data basis: representative rows as recorded, e.g. art curators hub avg = 30, growth.1m = 300; anna katz curator avg = 70, growth.3m = 550, growth.6m = 550). These are the most seductive rows in the dataset and the least reliable. A 550 percent gain on 70 searches is roughly 60 extra searches a month. No content or advertising decision should be anchored to that.

Risks & Limitations

This export carries seven distinct risks and limitations, and each one narrows what a reader can legitimately conclude from it.

Limitation 1 - no monthly history, therefore no seasonality judgement is possible. The most consequential limitation is the absence of any monthly time series. The run schema anticipates a history array of month-by-month search volumes for each topic, and none is present in any of the 1,000 rows (data basis: no trendHistory field in any topic object; counts.latestDataMonthTop = 2026-08 across all 1,000 topics). Every keyword in this export is anchored to a single month, August 2026. This limits three specific classes of judgement. First, seasonality cannot be assessed at all: a keyword that rises 80 percent in the most recent three months cannot be distinguished from one that rises 80 percent every autumn. Second, the difference between a structural trend and a one-off spike cannot be verified, because a spike and a trend look identical when you only have endpoints. Third, the direction of the seed term itself is only partly legible - curate shows minus 18.5 percent over one year and plus 21.5 percent over three years, but with only three data points and no intervening months, the shape of the recovery curve is unknowable (data basis: curate, growth = {1y -18.5, 2y 0, 3y +21.5}).

Limitation 2 - null growth fields at scale. Roughly 6 to 7 percent of rows have no growth value in any given short window: 57 rows missing one-month growth, 66 missing two-month, 62 missing three-month, and 70 missing six-month (data basis: summary.metrics.growth.1m.count = 943, 2m.count = 934, 3m.count = 938, 6m.count = 930, out of 1,000). The long windows are far worse - only one keyword in the entire export has one-year, two-year and three-year data (data basis: growth.1y.count = 1, growth.2y.count = 1, growth.3y.count = 1). This matters because it means no keyword except the seed can be classified as a durable multi-year trend. A reader cannot distinguish between a term that has been rising for three years and a term that appeared three weeks ago, for 999 of the 1,000 rows.

Limitation 3 - the competition index has a zero floor that hides inaction. More than half of the export sits at a competition index of exactly zero, and the median across all 1,000 rows is 0 (data basis: competitionIndex.p50 = 0, competitionIndex.avg = 17.1). A zero index is not a promise that a keyword is easy to rank for. It frequently means the keyword is so small that no advertiser has ever bothered to bid, and that the ad platform has no data to report. Recommending a zero-competition keyword as an opportunity without checking organic difficulty would be a category error, and this report has been careful throughout to treat zero-competition keywords as low-resistance but also low-value.

Limitation 4 - bid data covers only 18.2 percent of the export. Only 182 of 1,000 keywords carry any top-of-page bid data (data basis: lowTopOfPageBidMicros.count = 182). Every commercial conclusion in this report is therefore drawn from a minority of the dataset. The 818 rows with no bid data are not necessarily worthless - they may simply be too small for the platform to record a rate - but they cannot support a commercial thesis. This limitation is particularly acute for the museum-entity cluster, which is the largest cluster in the export and almost entirely bid-free.

Limitation 5 - brand and institutional trademark exposure. This is the sharpest legal and platform-compliance risk in the dataset. Dozens of keywords contain registered brand or institutional names, including Hooker Furniture and its Curata product line (roughly 19 keywords), Sotheby's, COSRX, Cerave, Bioderma, Garnier, Rilastil, Uriage, SVR, Guinot, Mary Cohr, Jonzac, Heimish, Proraso, Karcher, Rowenta, Swiffer, Decathlon, Ruris, Steinhaus, Wasserkonig, Wassertechnik, Samus, Daewoo, Chemstal, Bilka, Wolfson, hotel Aleph, and roughly forty museum and gallery names including MoMA, Tate, the V&A, the Met, the Whitney, the Guggenheim, the Smithsonian, LACMA, SFMOMA, the Rijksmuseum, the Centre Pompidou, the Louvre, Crystal Bridges, the Hammer, the Walker, the High, the Crocker, the Jewish Museum and the Natural History Museum, plus names of living curators such as Thelma Golden, Hans Ulrich Obrist, Michelle Grabner, Mami Kataoka, Kathy Halbreich, Rebecca Wilson, Cosmin Costinas and many more (data basis: keyword text across topics). Building content or advertising against a third party's brand name carries trademark and platform-policy exposure, and building content about a named living individual carries additional risk if the content is negative, comparative, or commercial.

Limitation 6 - medical and YMYL content exposure. The cure cluster contains a substantial number of health and treatment queries, including telogen effluvium cura, telogen effluvium si guarisce, come curare telogen effluvium, fibromialgia cura, cura asma, leishmania cura, dermatite canina cura, papillomavirus cura uomo, cycas revoluta cura, trichomonas cura antibiotica, cura narcisismo, cura agorafobia, curatarea ficatului de alcool, curatarea ficatului in mod natural, alimente pentru curatarea ficatului, dezintoxicarea ficatului, remedii naturale pentru curatarea colonului, spalare colon and iritatie in zona genitala dupa ras (data basis: keyword text across topics). Search engines apply additional quality standards to medical and health content, and the topics in this list range from genuine medical conditions to pseudoscientific detoxification claims. Any content built against these terms needs medical review, a clear authorship and sourcing standard, and a deliberate position on whether the query is being served or exploited.

Limitation 7 - the run is capped, the market is not. The run requested exactly 1,000 keywords and returned exactly 1,000, with 999 successfully expanded and zero failures (data basis: run.requestedCount = 1000, run.resultCount = 1000, run.checkedCount = 1000, run.expandedCount = 999, run.failedCount = 0). An exact match between requested and returned is a ceiling, not a census. This export cannot be used to estimate the total size of the curate keyword universe, and any statement of the form 'these are the keywords in this market' would be false. A further coverage restriction follows from the run configuration: the market is set to global with no geographic target list and no language filter beyond a single English constant, yet the results are visibly multilingual (data basis: run.marketKey = global, run.geoTargetConstants = empty, run.locale = en; and the presence of Romanian, Italian and Spanish keywords throughout). This means the export cannot be used to draw any conclusion about a specific country's behaviour, and any localisation decision needs a fresh, geo-targeted run.

The 2026-08 snapshot is a substantive constraint. All 1,000 rows share the latest data month of August 2026, and the export was produced in October 2026 (data basis: counts.latestDataMonthTop = 2026-08; run.exportedAt = 2026-10-11T05:40:49Z). The trend columns therefore describe a completed three-month window ending in August, not current conditions. Advertising or content decisions made on this data in late 2026 are being made on information that is at minimum two months old, and for a word as volatile as curate - where the museum-entity rows swing by hundreds of percent on tiny bases - two months is enough time for the pattern to have changed entirely.

Action Recommendations

The narrative arc of this report runs like this. The keyword universe for curate is enormous but structurally fragmented: one seed word has pulled in an art world, a set of non-English medical queries, and a set of non-English cleaning queries, all held together by nothing more than shared spelling (data basis: topics spanning curate, telogen effluvium cura and mașină de curățat cabluri under the same seed). Within that fragmentation, the art world dominates by row count but not by commercial value, because more than half of all rows sit at a competition index of zero and roughly 82 percent carry no advertiser bids at all (data basis: competitionIndex.p50 = 0; lowTopOfPageBidMicros.count = 182 of 1000). The one place where demand, advertiser willingness to pay, and low competitive resistance overlap is the curator career and credentials cluster, where bids run as high as 23.27 on single keywords and universities and course providers are visibly competing for prospective students (data basis: art curator education bidHigh = 23270000; museum and curatorial studies bidHigh = 20672674; curator graduate programs bidHigh = 20895514). The single largest genuine volume opportunity sits outside art entirely, in a medical query in Italian (data basis: telogen effluvium cura, avg = 12100, growth.3m = 83.5, bidHigh = 6374101), and it carries compliance risk that the career cluster does not. And the trend data, while noisy, is broadly pointing the same way: art-world celebrity and theory terms are declining while career and credential terms are rising (data basis: hans ulrich obrist growth.3m = -18.5, bourriaud growth.3m = -33.3, germano celant growth.3m = -28; against museum curator job growth.3m = +86.2, curator education requirements growth.3m = +66.7, museum curator training growth.3m = +100). The opportunity is therefore specific: build for the people who want the job, not for the people who want the gossip.

Content recommendations. The content strategy should be tiered by intent, and the first tier should be the career-and-credentials cluster. Build a comprehensive guide to becoming an art or museum curator, and build it to cover the exact queries in this export: how to become an art curator, how to become an art museum curator, how to become an art curator without a degree, how to become a curator of an art gallery, how do you become an art curator, becoming a museum curator, curator job requirements, curator education requirements, museum curator requirements, museum curator education requirements, curator degree, masters in curatorial studies, best curatorial studies programs, and museum curator course (data basis: all present in topics with growth between +23 and +100 percent and competition indexes between 4 and 59). This topic has a genuine advertiser market behind it, which means it can be monetised through affiliate or programme partnerships rather than through pure advertising. The second tier should be the definitional head terms, because they are enormous and nearly free to enter: what is an art curator at 1,000 searches with a competition index of 0, museum curator and curator in a museum at 9,900 each with a competition index of 1, and art curator at 6,600 with a competition index of 2 (data basis: all as recorded, with bid ranges from 0.68 to 8.53). These are the keywords that establish topical authority, and their bid data shows that advertisers think the audience is worth money even though the competition is thin. The third tier should be the theoretical and historical library content - the culture of curating and the curating of culture, thinking contemporary curating, relational aesthetics, documenta curators, curatorial statement example - but only as supporting material, because this tier is measurably shrinking (data basis: bourriaud growth.3m = -33.3; harald szeemann growth.3m = -18.7; germano celant growth.3m = -28).

What the content strategy should not do is chase the named-curator keywords on volume. A page about thelma golden art looks attractive at a score of 2,034.2, but the keyword has 50 monthly searches, its most recent month is down 65.6 percent, and it is a page about a living person (data basis: avg = 50, growth.1m = -65.6). The same is true of assistant curator v&a, v&a assistant curator, michelle grabner art and the other 400-plus museum-entity rows. If a site's topical authority genuinely covers a specific museum or a specific curatorial figure, a handful of these pages should exist; they should not form the spine of a content plan.

Product and sourcing recommendations. The most concrete product signal in this export is the Hooker Furniture Curata line. Roughly 19 keywords name this furniture collection directly, and several carry real advertiser bids despite negligible volume - curata writing desk at 1.63 to 6.28 and hooker curata dining table at 1.08 to 2.67 are the clearest examples (data basis: bidLow = 1628041, bidHigh = 6281088; and bidLow = 1083427, bidHigh = 2674365). For a furniture retailer or a home-goods marketplace, the implication is straightforward: curata is a live product-line search term, demand exists, and the competition index of 100 on these keywords indicates that authorised and unauthorised sellers are already bidding (data basis: curata desk competitionIndex = 100; curata dining table competitionIndex = 100; curata nightstand competitionIndex = 83). Any sourcing decision here should be made through an authorised dealer relationship rather than around the trademark. A second product signal is in skincare: cerave acid salicilic, cosrx acid salicilic and bioderma node a sampon all show competition indexes of 100 with volumes between 260 and 590, and cosrx acid salicilic additionally shows 323.5 percent six-month growth (data basis: cosrx acid salicilic avg = 260, competitionIndex = 100, growth.6m = 323.5). That is a market where demand is real but brand-dominant; a private-label or adjacent-ingredient play (rather than a direct brand play) is the only realistic route in.

The third product signal, and the one with the least obvious mapping, is the educational product. The bid data on museum and curatorial studies (up to 20.67), curator graduate programs (up to 20.90), masters in museum curation (up to 20.64) and art curator education (up to 23.27) tells you that someone is paying serious money to reach people who want to become curators (data basis: all bidHigh values as cited). If the business has any educational capability - a course, a certification, a mentorship programme, a portfolio-review service - this is the single best-signalled opportunity in the export, because the advertiser willingness to pay is documented and the organic competition is minimal.

Ad spend recommendations. For paid search, the bid data in this export is unusually informative because it tells you where the market is willing to pay and where it is not. The highest-value, lowest-competition targets are the software and event terms: curate event software at 4.60 to 19.77 and curate wedding software at 3.10 to 40.32 both sit on 10 monthly searches with competition indexes of 69 and 43 respectively (data basis: bidLow = 4600000, bidHigh = 19765532; bidLow = 3100000, bidHigh = 40315260). These are B2B software terms where the customer lifetime value justifies extraordinary cost per click; the low volume means a modest budget can dominate the auction entirely. The second highest-value target set is the education cluster, where art curator education, museum and curatorial studies, curator graduate programs and best colleges for museum curators all show bid ceilings above 14.00 (data basis: bidHigh values of 23270000, 20672674, 20895514 and 14502554). The third target is the general career head terms: museum curator job at 0.07 to 2.00 with a competition index of 2, and curator job requirements at 0.28 to 2.00 with a competition index of 1, are cheap to enter precisely because almost nobody is competing (data basis: both as recorded).

What should not be bought is equally clear. The Romanian cleaning and hardware keywords are the worst value in the export: they combine a competition index of 100 with volumes of 10 to 30 and bid ranges as low as 0.04 to 0.11 (data basis: spray de tastatura avg = 30, competitionIndex = 100, bidLow = 37691, bidHigh = 110980). Paying to enter a saturated auction on a keyword nobody searches is a straightforward loss. Equally, the museum-entity long tail should not be bought: with no bid data recorded on the vast majority of these terms and volumes of 10 to 50, there is no auction to win (data basis: tate curator avg = 30, no bid fields; met museum curator avg = 50, no bid fields; louvre curator avg = 110, no bid fields). And the medical cluster should be entered only if the business is genuinely a healthcare provider or a licensed content publisher, at which point the priority target is telogen effluvium cura - but only after verifying its language scope, its seasonality (which this export cannot establish) and its regulatory position (data basis: avg = 12100, competitionIndex = 52, growth.3m = 83.5, no monthly history supplied).

Three immediate next steps, in order. First, run a fresh, geo-targeted and language-restricted export on curate in English-only, US or UK, to strip out the Romanian cleaning keywords and the Italian medical keywords that are inflating this dataset by roughly a quarter of its rows; that single change would turn a fragmented export into a usable art-market map (data basis: run.geoTargetConstants was empty and run.locale was a single English constant, yet the export contains substantial Romanian, Italian and Spanish content). Second, verify the education cluster's values independently, because the bid ceilings of 20 to 23 currency units on 10-to-50-search keywords are extreme enough that a data error is possible and the entire monetisation thesis rests on them (data basis: art curator education bidHigh = 23270000 on avg = 10; curator graduate programs bidHigh = 20895514 on avg = 10). Third, request a monthly time series for the top 50 keywords in this list, because without it no seasonality claim, no spike-versus-trend distinction, and no durable-versus-fleeting judgement can be made - and roughly a third of the rising keywords in this export are rising only on bases small enough that the signal may not survive that test (data basis: no trendHistory present in any of the 1,000 topics; growth.1m.max = 325, growth.2m.max = 1000, growth.3m.max = 1000, all appearing almost exclusively on keywords with avgMonthlySearches below 70).