Native Advertising Statistics 2026: A July Snapshot of 1,041,637 Captured Ads
A July 2026 snapshot of our own collection, not a survey: 1,041,637 creative records, 36,009 resolved advertisers and 10.75M observations across 50 networks, with per-network counts, category mix and observation spans.

Most native advertising statistics pages recycle the same decade-old survey findings ("70% of people prefer to learn about products through content"). This page reports a dated snapshot of OpenAdLibrary's own ad collection instead. As of late July 2026, the collection covered 50 native ad networks, gathered mainly through API requests with a browser crawler as a fallback, and held 1,041,637 creative records from 36,009 resolved advertisers, with 1,975,205 landing-page captures and 10,751,604 ad observations behind them. These are counts of collected records you can inspect individually, not survey estimates and not totals for the market. Two third-party market estimates are labelled separately.
If you cite a figure, attribute it as "OpenAdLibrary, July 2026 collection snapshot" and keep its caveats: it describes this collection's coverage, not market size, spend or performance. A methodology section with caveats is at the end; the macro narrative lives in our companion State of Native Advertising 2026 report.
The page covers headline totals, per-network record counts, the classified category mix, observation spans, the longest spans in the snapshot and capture depth, followed by two third-party market-size estimates for context. The first-party figures are counts from one vendor's collection; they are not a census of native advertising.
Headline statistics#
OpenAdLibrary's collection held 1,041,637 native ad creative records in the July 2026 snapshot (OpenAdLibrary, July 2026). 615,327 records, 59.1% of the total, were observed at least once in the 30 days before the snapshot (OpenAdLibrary, July 2026). 489,511 records were first observed in those 30 days, about 16,300 per day; first observation is not a launch date, and collection changes can raise this count (OpenAdLibrary, July 2026). The records were attributed to 36,009 resolved advertisers, an average of about 29 records per advertiser (OpenAdLibrary, July 2026). The collection covered 50 native ad networks, including Taboola, Outbrain, the Microsoft Audience Network, MGID, Revcontent, MediaGo and Yahoo (OpenAdLibrary, July 2026). 1,975,205 landing pages had been captured by following ad click chains through redirects, about 1.9 per creative record (OpenAdLibrary, July 2026). 10,751,604 ad observations (sightings of a creative in a collected feed) back the snapshot, an average of 10.3 per record and the raw material for the observation spans below (OpenAdLibrary, July 2026).
| Metric | Value (July 2026 snapshot) |
|---|---|
| Creative records collected | 1,041,637 |
| Records observed in prior 30 days | 615,327 |
| Records first observed in prior 30 days | 489,511 (~16,300/day) |
| Resolved advertisers | 36,009 |
| Networks covered | 50 |
| Landing-page captures | 1,975,205 |
| Ad observations | 10,751,604 |
Captured records by network#
Taboola had the most creative records in the snapshot: 352,392 collected, 254,020 of them observed in the prior 30 days, a 72.1% share (OpenAdLibrary, July 2026). The Microsoft Audience Network was second on record count with 332,883, but only 132,784 (39.9%) were observed in the prior 30 days, the lowest share among the majors. That share depends on how often and where each feed was collected as well as on how quickly advertisers change creatives, so it does not measure feed turnover on its own. (Deep dive: MSN native ads guide.) Outbrain (now Teads) accounted for 150,462 creative records, 91,668 observed in the prior 30 days (OpenAdLibrary, July 2026). MGID accounted for 96,139 records, 59,604 observed in the prior 30 days (OpenAdLibrary, July 2026). Revcontent accounted for 23,040 records, 15,891 observed in the prior 30 days (OpenAdLibrary, July 2026). Yahoo native accounted for 14,389 records, 12,473 observed in the prior 30 days, an 86.7% share; that is the highest share among the majors in this collection, not proof of the most stable creative pool (OpenAdLibrary, July 2026). MediaGo, which describes itself as a deep-learning advertising platform (MediaGo DSP), accounted for 7,002 records, 797 observed in the prior 30 days (OpenAdLibrary, July 2026). Taboola and Microsoft together held 685,275 records, 65.8% of the collection, and adding Outbrain brings the top three to 80.2% (OpenAdLibrary, July 2026). These are shares of this collection, not market shares.
| Network | Creative records | Observed in prior 30 days | Share observed in prior 30 days |
|---|---|---|---|
| Taboola | 352,392 | 254,020 | 72.1% |
| Microsoft Audience Network | 332,883 | 132,784 | 39.9% |
| Outbrain (Teads) | 150,462 | 91,668 | 60.9% |
| MGID | 96,139 | 59,604 | 62.0% |
| Revcontent | 23,040 | 15,891 | 69.0% |
| Yahoo | 14,389 | 12,473 | 86.7% |
| MediaGo | 7,002 | 797 | 11.4% |
The remaining 43 monitored networks — long-tail exchanges and regional players — account for the balance of the index (65,330 tracked creatives between them).
Category mix of classified records#
Category shares below are of classified records only: 182,082 records carried a category classification in the July 2026 snapshot, about 17.5% of all records. The unclassified majority may have a different mix.
Health was the largest classified category, at 14.1% of classified records (OpenAdLibrary, July 2026). Insurance was second at 10.8% and finance third at 10.2%, together 21.0% of classified records (OpenAdLibrary, July 2026). (Examples: finance native ads.) On Taboola, health was 19.2% of the network's 82,142 classified records, with insurance at 11.4% and finance at 10.0%, together about two in five classified Taboola records (OpenAdLibrary, July 2026). These shares describe classified records in this collection, not spending or market share.
These figures cannot settle claims that one category "owns" native. Supplement and other nutra offers may be classified under health, only about 17.5% of records are classified, and record counts are not spend. We discuss the category data further in our vertical deep-dive.
| Category | Share of classified records (collection-wide) | Share of classified records (Taboola) |
|---|---|---|
| Health | 14.1% | 19.2% |
| Insurance | 10.8% | 11.4% |
| Finance | 10.2% | 10.0% |
Observation spans: how long creatives were seen#
These are the most misunderstood numbers in native. We record first-seen and last-seen dates for every creative record, and most records were seen only briefly. A short observation span is not the same as a short campaign:
59% of creative records were not observed again after the day they were first observed (OpenAdLibrary, July 2026). 25.2% were observed across at least 7 days (OpenAdLibrary, July 2026). 10.7% were observed across at least 30 days (OpenAdLibrary, July 2026). The snapshot reported an average observed span of 7.1 days for creatives it called "matured" without defining that group, so treat that figure as unverified. These are capture statistics, not lifespans or a death rate: a creative can keep serving where the collector did not look, and about 47% of records were first observed in the final 30 days, so many could not yet show long spans. Long spans are uncommon here, but they do not show profitability; brand campaigns, testing budgets, slow optimization and capture gaps can all produce them, and spend and conversions are not observed.
| Observation milestone | Share of creative records |
|---|---|
| Observed again after the first observed day | 41.0% |
| Observed across at least 7 days | 25.2% |
| Observed across at least 30 days | 10.7% |
Important caveat: the tracking window for this snapshot opened on June 2, 2026, so observation spans are bounded by the window. A span is the time between first and last observation, not proof of continuous delivery. Creatives that started before June 2 have understated spans, and creatives first observed close to the snapshot date could not yet show long ones. Capture cadence and coverage also affect whether a live creative is seen again, so the short-span shares are not a measured death rate.
Longest observation spans in the July 2026 snapshot#
Because the window opened June 2, 2026, the longest possible span at this snapshot was roughly eight weeks, and several networks had creatives observed across nearly all of it:
In the July 2026 snapshot, the longest observation spans on the Microsoft Audience Network and Outbrain were at least 56 days, and Taboola, MGID, Revcontent and MediaGo each had creatives observed across at least 55 days, close to the full window (OpenAdLibrary, July 2026). Creatives first observed on June 2 may have started earlier. These spans show repeated observation, not continuous delivery, spend or profitability; we list examples in the longest-running native ads.
Advertisers and the long tail#
With 36,009 resolved advertisers behind 1,041,637 creative records, the average is about 29 records per advertiser; the snapshot does not report the median or how concentrated records are among advertisers. Grouping records by resolved advertiser rather than by the brand label on the card shows operators such as comparison sites, search-arbitrage buyers and insurance lead generators appearing across several networks alongside consumer brands. Advertiser resolution can be uncertain, and record counts are not spend. The full rankings and per-advertiser drill-downs live in top native advertisers by network and who advertises on Taboola.
Capture depth#
Each creative record is backed by an average of 10.3 feed observations (10,751,604 observations across 1,041,637 records), the repeated sightings that make first-seen and last-seen observation spans possible (OpenAdLibrary, July 2026). Landing pages outnumber creatives: 1,975,205 landing captures against 1,041,637 creative records (about 1.9 per record), which can happen when one creative rotates through several pre-landers and tracker destinations (OpenAdLibrary, July 2026). Captured landing pages also feed our copycat landing-page detection, which flags pages that substantially copy another brand's page.
Industry context (external sources)#
For sizing the market around these observations, two third-party forecasts are worth citing alongside our capture data:
US native display ad spending is forecast to grow 13.1% in 2026 to $147.98 billion, per eMarketer's December 2025 forecast. Future Market Insights estimates the global native advertising market at $125.6 billion (USD 125,589.1 million) in 2026 in its 2026-2036 outlook published February 9, 2026, and Market Research Future models $124.88 billion for 2025 growing at a 12.37% CAGR to 2035 (page updated September 17, 2026). Because a US-only forecast exceeds both global estimates, the sources cannot share a definition of native advertising; check each source's scope before citing them together.
We treat external market-size figures as context only; every other number on this page is first-party capture data.
What the July 2026 snapshot can and cannot show#
Five takeaways, with their limits:
Most records were seen briefly and a small share were seen for a long time. Neither fact shows how long campaigns ran or which angles worked, because long spans can come from brand campaigns, test budgets, slow optimization or capture timing. Two networks account for most of this collection. Taboola and the Microsoft Audience Network held 65.8% of captured creative records, which reflects collection coverage and cadence, not market share. Microsoft does publish an Ad Library, but its documentation describes it as covering ads shown on Bing.com and served in the European Economic Area; it does not describe coverage of the native placements counted here. Health, insurance and finance were the largest classified categories. They made up 14.1%, 10.8% and 10.2% of the roughly 17.5% of records that carry a classification; that does not show where spending concentrates or why. The collection includes many non-English creatives. It samples several geographies and device profiles, so researchers can find examples outside English-language markets (see scaling into new geos). Counts change quickly. About 47% of records were first observed in the 30 days before the snapshot, which partly reflects the window opening on June 2, 2026 and changes in collection, so treat this page as a dated snapshot rather than a current census.
For media buyers, the practical use is narrower: use the snapshot to find examples to inspect, compare networks using their current documentation and your own tests, and judge competitors by what you can verify, not by record counts or observation spans.
Methodology#
What OpenAdLibrary captures. OpenAdLibrary collects native ad creatives (image, headline, brand label), placement observations and click-chain data from native advertising networks (50 in the July 2026 snapshot), including Taboola, Outbrain/Teads, the Microsoft Audience Network, MGID, Revcontent, MediaGo and Yahoo. Collection is API-first: for native networks we request the same ad feeds that publisher pages request, across desktop, Android and iOS device profiles and multiple geos, with a browser crawler as a fallback. Click chains are followed through tracker redirects, without clicking live ads, to identify the likely advertiser and landing page; landing pages are captured and stored (1,975,205 captures in the July 2026 snapshot).
Snapshot and windows. All figures are a July 2026 snapshot, refreshed in late July. The current tracking window opened on June 2, 2026; "live" counts creatives observed in the trailing 30 days, and all longevity figures are therefore "at least N days" — lower bounds, not exact lifespans. Advertiser identity is resolved by unwrapping tracker redirects and deduplicating brand labels, so "36,009 advertisers" counts entities, not display names.
Caveats. (1) Creative counts measure creative production as captured, not spend; high-rotation advertisers weigh more than steady ones. (2) The tracking window began June 2, 2026, so long spans are understated and recently first-seen records cannot yet show long spans. (3) Category shares are of classified records only: 182,082 records carry a classification (assigned from traced landing pages, resolved advertisers, copy and imagery), about 17.5% of the collection. (4) Geo coverage follows our monitored markets and expands over time; per-network shares can shift as coverage does. (5) Per-network shares observed in the prior 30 days depend on each network's capture cadence as well as real feed turnover. (6) First and last observation dates are not launch or end dates, and a span is not continuous delivery. (7) None of these counts measures spend, clicks, conversions or profitability.
Verification. These figures come from a July 2026 snapshot of a changing collection, and the queries behind them were not published with this page, so the product will not reproduce them exactly today. Founders subscribers ($29.99/month) can search the library with network, category and longevity filters; the free plan (no card) shows the 20 newest captured Meta, TikTok and native ads. The individual creatives cited across our example galleries, Taboola and finance, can be inspected the same way. Journalists and researchers: cite as "OpenAdLibrary, July 2026 collection snapshot" and link this page; programmatic access is through our native ad data API on the Founders plan. This page was last updated in late July 2026, so check the date before reusing a figure.







