Bannerflow on September 14, 2026 published an analysis of anonymised data from more than 300 brands on its creative platform, finding that 78% had produced video advertising between January and August 2026, while only 46% had served video impressions over the same period.

In Short

Bannerflow looked at more than 300 companies using its ad-making software this year and found that most of them made video ads, but fewer than half actually showed any video to people. That matters because a video ad costs time and money to make, and one that never runs does nothing for the company that paid for it. The numbers suggest the harder part of video advertising is no longer making the ads but getting them into live campaigns - and judging them by the right yardstick once they are running.

What the dataset covers

The figures come from brands active on Bannerflow's platform between January 1 and August 31, 2026, a window of 243 days. Over that period, according to Bannerflow, those brands produced 2.33 million creative assets, served 155.1 billion ad impressions and recorded 166.8 million clicks. Divided across the window, that is roughly 638 million impressions and about 686,000 clicks a day.

The headline finding is a gap of 32 percentage points between making video and running it. According to Bannerflow, 78% of the brands in the sample produced video in 2026, but only 46% served video impressions. Because the company describes the sample only as "300+" brands, the absolute numbers can be estimated solely as lower bounds: at least 234 brands produced video, and at least 138 served it.

Bannerflow frames this as an activation gap. The press materials describe the industry's next video problem as deployment rather than production, arguing that competitive advantage is moving towards distribution and the ability to reach the right audiences quickly. That is the company's interpretation of its own data, and the release contains little that tests it directly.

The two percentages also measure slightly different things. A brand counts towards the 78% if it produced any video during 2026, and towards the 46% if it served any video impressions. Nothing in the release states that the second group is a subset of the first. A brand could, in principle, have served video built in 2025 without producing anything new this year, which would mean the number of brands that made video and left it unused is not simply the difference between the two figures.

Broad but shallow: video's share of output

The second finding qualifies the first. Although 78% of brands produced some video, the format accounted for only 10.5% of all creative assets, according to Bannerflow. Static creative represented the remaining 89.5%.

Applied to the 2.33 million total, those shares translate to roughly 244,650 video assets against about 2.09 million static ones - approximately eight static assets for every video. The company summarises this as adoption that is "broad but still shallow."

Asset counts carry a structural bias, however. A single static display campaign typically generates many assets from one idea, because each ad size, language, currency and market variant is counted separately. PPC Land's report on SEGA's use of Bannerflow across seven markets on August 12, 2026 described exactly that multiplication: language, currency, age rating and size requirements across the United States, United Kingdom, France, Germany, Italy, Spain and Australia pushing a single campaign's asset count into the hundreds. Video tends not to multiply in the same way. An asset-share figure therefore says more about the volume of production work than about how much media weight video carries.

The release does not disclose what proportion of the 155.1 billion impressions were video. That figure would show whether video is marginal in delivery or only in asset counts, and its absence leaves the central claim partly open.

iGaming dominates the impression base

Sector data forms the most technically revealing part of the analysis. According to Bannerflow, iGaming accounted for 76.1% of all impressions in the dataset but 63.8% of clicks. Travel showed the reverse pattern, generating 11.0% of clicks from 4.3% of impressions.

Those shares imply absolute volumes. Three-quarters of 155.1 billion impressions puts iGaming at roughly 118 billion impressions and about 106 million clicks. Travel, by the same arithmetic, accounts for around 6.7 billion impressions and 18.3 million clicks.

The consequence for interpretation is considerable. With more than three in four impressions coming from one vertical, the platform's overall figures are largely iGaming figures. The blended click-through rate of 0.108% that Bannerflow uses as its benchmark - a number consistent with dividing 166.8 million clicks by 155.1 billion impressions - is pulled down by iGaming's own rate, which the shares imply sits at about 0.090%. Every other sector combined, telecom included, accounts for 19.6% of impressions and 25.2% of clicks, an implied rate of roughly 0.138%.

When the release says travel outperformed "the platform rate," then, it is comparing travel mostly against iGaming. The release does not disclose how many brands operate in each vertical, so it is not possible to tell whether iGaming's weight reflects a handful of very large advertisers or a broad cluster of smaller ones.

That concentration has a regulatory dimension. Gambling advertising runs under country-by-country restrictions that shape which inventory can be bought and how. Google, for example, removed its certification requirement for gambling creatives in 37 markets on Authorized Buyers from August 10, 2026 - inside Bannerflow's measurement window - while on the direct-buy side it had tightened enforcement against repeat violators and manager accounts in January. A dataset in which one regulated category supplies most of the volume will reflect that category's buying constraints as much as any general creative trend.

Telecom and travel: what the video figures do and do not show

Bannerflow pairs its sector click data with video shares. According to the company, telecom, where 24.1% of impressions were video, recorded a blended click-through rate of 0.208%, nearly twice the platform rate of 0.108%. Travel, where 12.5% of impressions were video, achieved the highest click-through rate in the dataset at 0.274%, more than 2.5 times the platform rate.

The ratios check out: 0.208% is 1.93 times 0.108%, and 0.274% is 2.54 times it. Travel's rate can also be reproduced independently from the share figures. Eleven per cent of clicks divided by 4.3% of impressions, applied to the platform average, gives about 0.275%.

What the numbers cannot establish is that video drove those rates. Travel had a lower video share than telecom yet a higher click-through rate, which runs against any simple reading that more video produces more clicks. The release also does not break out click-through rates for video and static impressions separately, nor does it give iGaming's video share. Both telecom and travel figures are blended, meaning video and static delivery are mixed in one number.

Clicks are also an imperfect measure for video. Many video placements are bought for reach or completed views rather than for traffic, and a campaign on connected television can deliver its value without producing a click at all. Bannerflow's own spokesperson made a version of this point, noting that broad-reach and high-intent campaigns do different jobs.

What Bannerflow says

Jamie Day, Head of Marketing and Lead Generation at Bannerflow, set out the company's reading of the data. "Marketers have solved the first half of the video problem: making the assets. But production is not the same as adoption. If 78% of brands are producing video but only 46% are actually serving it, then output alone is no longer a meaningful measure of whether a video strategy is working," Day said.

He continued: "The opportunity is to connect creative and media workflows more closely, plan distribution alongside production, and put existing video assets to work for the right audiences. Campaigns must also be assessed against their objective: broad-reach and high-intent campaigns do different jobs, so reach and click-through rate need to be interpreted in context."

The second half of that statement sits in some tension with the release's own presentation, which leads its sector findings with click-through rate comparisons. The dataset as published does not classify campaigns by objective, so readers cannot see whether travel's higher click rate reflects a greater share of performance-focused campaigns rather than anything about the creative.

The production boom behind the gap

The finding lands after roughly eighteen months in which platforms and vendors concentrated on making video cheaper to produce. Research from the Interactive Advertising Bureau in July 2025 found that 86% of buyers use or plan to use generative AI to build video ads, with buyers projecting that generative AI creative would account for 40% of all advertisements by 2026.

The tools followed. Google brought its Veo video model into Asset Studio, its creative workspace inside Google Ads, in 2025, generating video from static images and text prompts. Amazon extended Video Generator to all United States advertisers on June 10, 2025, as PPC Land noted in its coverage of Amazon's interactive video format for Sponsored Products. Microsoft added Image Animation to its Copilot creative tools on November 17, 2025, converting static images into video assets for its advertising network. By mid-2026, a weekly roundup of industry developments was citing figures showing 91% of advertisers using or planning to use AI for creative production, and Cadbury generating 130,000 unique AI videos for a single campaign.

Not every survey agrees on how far that shift has gone. TripleLift's research in May 2026 found that 73% of advertising professionals used AI for optimisation but only 25% for creative production, with 67% citing lack of trust in AI output as a reason for manual oversight. Different samples and definitions explain much of the variation, but the direction is consistent: the cost of producing a video asset has fallen, and the constraint has moved elsewhere.

Where exactly? Bannerflow's answer is distribution. Other evidence points to adaptation. A survey by Smartly and EMARKETER, cited in PPC Land's report on common mistakes in connected TV campaigns, found that 72% of marketers reused or only slightly modified assets across social media and connected TV, while 25% tailored creative for both. An asset built for one surface does not automatically fit another, and a video produced for a social feed may never be trafficked to a television screen at all.

Variant volume appears to matter on some platforms. LinkedIn reported in July 2026 that advertisers running five or more ad variants saw 20% higher click-through rates. Amazon, meanwhile, has moved variant generation into delivery itself: its Dynamic TV Creative product for Prime Video, covered in May 2026, generates versions at the moment of impression from one base asset. Platforms that assemble creative at serve time remove the separate production step from the equation entirely, which may change what an "unserved" asset even means.

Who Bannerflow is

Bannerflow was founded in Stockholm in 2010. The company says it has surpassed 25 million euros in annual recurring revenue and employs more than 135 people across five international offices. According to Bannerflow, more than 2,000 brands use the platform, including Boozt, SAS, Ooni, SEGA and Meliá Hotels International.

The platform combines creative production, automation, personalisation, activation, intelligence and measurement, according to the company, with campaigns managed across display, social, video, CTV, DOOH and other digital channels from one interface. The release describes Bannerflow in two slightly different ways - as an "AI-powered creative automation platform" in its opening and as an "AI-native Creative Intelligence platform" in its notes to editors - a shift in positioning language rather than a substantive discrepancy.

The company also states it is recognised as a category leader by G2 and Forrester. The most recent Forrester evaluation in PPC Land's archive, the Q4 2020 assessment of creative advertising technologies, placed Bannerflow among strong performers, with Innovid, Jivox, Clinch and Flashtalking named leaders; Bannerflow was cited alongside Celtra and RevJet for workflow and collaboration tools. The release does not specify which Forrester or G2 assessment it refers to.

SEGA's case, published a month earlier, shows the kind of claim the company attaches to its product. SEGA reported cutting production time by two thirds and recording a 44% lower cost per thousand impressions and 14% lower cost per click on a dynamic social format built on Bannerflow. Those were customer-reported results against the customer's own baseline, and like the current dataset they were not independently audited.

Limits of the analysis

Several caveats apply beyond those already noted. The dataset is Bannerflow's own, drawn from its customers, and has not been verified by a third party. The sample of more than 300 brands represents roughly 15% of the 2,000-plus brands the company says use its platform, and the release does not explain how "active" brands were selected or whether the sample is weighted by spend.

A more basic question concerns where impressions are counted. A creative platform records the impressions it serves or tracks. If a brand produces a video in Bannerflow and then uploads it directly to a social platform or a video network that serves it natively, that delivery may not appear in Bannerflow's impression data at all. The release does not say whether exported video is captured. If it is not, some of the 32-point gap would reflect measurement boundaries rather than unused assets.

Finally, the analysis covers a period of eight months, without a comparison to 2025. Whether the gap is widening, narrowing or stable is not addressed.

Why it matters for the marketing community

For media buyers and creative teams, the Bannerflow data puts a number on a pattern that has been discussed mostly in anecdote: video production capacity has grown faster than the media plans built to use it. The finding matters because the economics of production have changed. When a video asset took weeks and an agency fee to make, producing it without a placement was a visible waste. When generative tools produce video from a still image in minutes, idle assets become cheaper to create and easier to overlook.

The sector data carries a second lesson, one about benchmarks rather than video. A platform-wide average built on a dataset where 76.1% of impressions come from iGaming is not a neutral comparison point for a travel brand, a telecom operator or a retailer. PPC Land has tracked efforts to bring more external reference points to engagement metrics, including BIScience's addition of panel-based click-through rate benchmarking to AdClarity in December 2025. Vendor-reported averages remain common in the market, and their composition often goes undisclosed.

Bannerflow's release is, in the end, a company describing a problem its product is designed to address. The data it published is more detailed than many such releases, and the arithmetic holds together. The open question is the one the release does not answer: how much of the 32-point gap is creative that never found a placement, and how much is creative that ran somewhere Bannerflow could not see.

Timeline

Summary

Who: Bannerflow, a Stockholm-founded creative platform company reporting more than 25 million euros in annual recurring revenue and more than 2,000 client brands, analysed anonymised data from more than 300 brands active on its platform. Jamie Day, its Head of Marketing and Lead Generation, commented on the findings.

What: The analysis found that 78% of brands produced video in 2026 but only 46% served video impressions. Video made up 10.5% of 2.33 million creative assets. iGaming supplied 76.1% of 155.1 billion impressions and 63.8% of 166.8 million clicks, while travel recorded the highest blended click-through rate at 0.274% against a platform rate of 0.108%, and telecom reached 0.208%.

When: The data covers January 1 to August 31, 2026. Bannerflow published the findings on September 14, 2026.

Where: The data reflects campaigns run through Bannerflow's platform, which spans display, social, video, connected TV and digital out-of-home channels across international markets. The release does not break results down by country.

Why: Bannerflow argues that falling video production costs have shifted the challenge from making video to deploying it. The figures matter for advertisers and agencies because they suggest a share of produced video never reaches audiences, and because a platform average dominated by one vertical is a weak benchmark for others - though the data is vendor-reported, unaudited and may not capture video served outside the platform.