Creative Advertising Technologies

Digital advertising spent twenty years optimising everything except the advertisement.

Targeting became a science. Bidding became an auction with millisecond decisions. Placement, frequency, attribution, and audience modelling all acquired dedicated technology categories and specialist vendors. The creative itself, the thing the human being actually sees, stayed roughly where it had been since television: made by people with taste, approved by committee, and shipped in a handful of versions.

Creative advertising technology is the attempt to close that gap. It treats the advertisement as a variable rather than a fixed input, which is a considerably more disruptive idea than it sounds, because it puts a system where a creative director used to stand.

What the technology does

The core mechanism is dynamic creative optimisation. Rather than producing one advertisement, you produce a template with variable elements: headline, image, product, offer, call to action, background, price. The system assembles a version at the moment of serving, using whatever data is available about the person, the context, the time, or the inventory.

Around that sit several other functions. Creative management, meaning the production and versioning of assets at volume. Ad serving and trafficking, getting the right file to the right placement in the right format. Personalisation logic, deciding which variant serves to whom. Testing and measurement, working out which variants actually performed. And increasingly generation, where the variants are produced by a model rather than assembled from prepared components.

Forrester's evaluations of this market have consistently been addressed to B2C marketing professionals rather than to technologists, which is a signal in itself. This is bought by marketing, on marketing's logic, and evaluated against marketing outcomes.

The vendor list is a graveyard

Three Forrester evaluations exist, and reading them in sequence is the most informative thing available about this category.

The Forrester Wave: Creative Advertising Technologies, Q4 2018, authored by Joanna O'Connell with Susan Bidel, scored ten providers against thirty one criteria: Adacado Technologies, Adform, Celtra, Clinch, Flashtalking, Google, Jivox, RevJet, Sizmek, and Thunder.

The Forrester Wave: Creative Advertising Technologies, Q4 2020, also by O'Connell, scored nine against thirty seven criteria: Adacado, Bannerflow, Celtra, Clinch, Flashtalking, Innovid, Jivox, RevJet, and SundaySky.

The Forrester Wave: Creative Advertising Technologies, Q4 2024 scored nine against thirty five criteria, with Smartly and Flashtalking both placing as Leaders.

Track what happened between editions. Google appeared in 2018 and was gone by 2020. Sizmek and Thunder both disappeared. Flashtalking now trades as Flashtalking by Mediaocean following acquisition. The vendor count went ten, nine, nine, while the composition kept turning over.

Two things are worth drawing out of that.

The first is that Google's presence and then absence is unusual. A company with that much advertising infrastructure appearing in a specialist creative evaluation and then not appearing suggests either a strategic retreat from the specialist positioning or a decision not to participate. Either way, the largest player in digital advertising is not competing in this category as a specialist.

The second is that this has been a hard market to survive in as an independent. The pattern of disappearance and absorption is consistent, and it points at the structural problem underneath: creative adtech sits between the creative agencies who make advertisements and the media platforms who serve them, and both of those neighbours have been expanding into the middle.

Why creative became the variable that mattered

The strategic case for this category strengthened considerably for a reason that had nothing to do with creative.

As third-party cookies degraded and identity-based targeting became less reliable, the levers available to an advertiser narrowed. If you cannot target as precisely, the thing you can still control is what the advertisement says and shows. Flashtalking's own framing of its Leader placement put it directly: in a world rife with AI and devoid of cookies, creative is emerging as the most critical variable for maximising return on ad spend.

That is vendor language and it is also a fair reading of the situation. When targeting precision falls, creative relevance has to do more work. An advertisement that adapts to context can partially substitute for an advertisement delivered to a precisely known individual.

There is a second argument that predates the cookie problem and has more evidence behind it. Industry research on advertising effectiveness has consistently found that creative quality accounts for a larger share of variance in campaign outcomes than media targeting does. If that is true, then the industry's allocation of technology investment has been backwards for most of the last decade, with enormous sophistication applied to the smaller lever.

Creative adtech is the correction. Whether it is a sufficient correction is a separate question, because optimising creative and making good creative are not the same activity.

The tension nobody in the category resolves

Here is the argument that runs underneath every product in this market, and it does not have a clean answer.

Dynamic creative optimisation works by decomposing an advertisement into interchangeable parts. That decomposition is what makes variation cheap. It also constrains what the advertisement can be, because anything that depends on the whole holding together, a joke, a narrative, a visual idea that only works in one arrangement, cannot survive being split into slots.

So the technology is excellent at the advertising that is fundamentally informational. Product, price, availability, offer, proximity. It is much weaker at the advertising that builds brands, because brand-building creative relies on exactly the coherence that templating removes.

Most large advertisers need both. The performance half benefits enormously from this technology. The brand half is where the marketing organisation's most senior and most protective people live, and their objection to creative adtech is not technophobia. It is a correct observation that a system optimising click-through rate will reliably converge on advertising that looks like everyone else's.

This is the actual reason adoption in this category has been uneven despite a compelling efficiency argument. It is not an integration problem. It is a disagreement inside the marketing organisation about what advertising is for, and buying a platform does not settle it.

Vendors that handle this well distinguish clearly between the workflows, giving performance teams the assembly-line capability and giving brand teams control over what can and cannot be varied. Vendors that handle it badly present one pipeline and let the argument play out during rollout.

Generation changes the constraint again

The current phase of this market moves from assembling variants to generating them.

Forrester's 2024 evaluation covered the full creative lifecycle from ideation through performance tracking, and Smartly's Leader positioning centred on AI-powered creative and campaign innovation integrating creative and media workflows, particularly across paid social and shoppable formats.

The shift is meaningful. Assembly required someone to produce every component in advance, so the variant space was bounded by production capacity. Generation removes that bound. A model can produce a thousand headline variations, restyle an image for a dozen placements, or resize and recompose a video for six formats without a human touching any of them.

Which relocates the constraint rather than removing it.

When variants were expensive, the hard question was which ones to make. When variants are free, the hard question becomes which ones to run, and that is a measurement problem. A thousand variants split across a campaign budget produces a thousand samples too small to distinguish from noise. The statistical difficulty of learning anything from a large variant space is real, and it is the quiet reason many creative optimisation programmes report inconclusive results.

The second constraint is brand consistency, which becomes considerably harder when the organisation is no longer approving every asset that runs. If a model generates copy and imagery at volume, brand governance has to move from reviewing outputs to constraining generation, and most brand functions have neither the tooling nor the vocabulary for that yet.

And the third is legal exposure. Generated imagery raises questions about rights, likeness, and provenance that a stock photography licence used to settle. An advertisement is a published claim, and the organisation publishing it carries the liability regardless of what produced it.

Where this sits against adjacent categories

Creative adtech overlaps with two neighbours, and the boundaries have been moving.

Digital asset management holds the approved assets, the brand rules, and the rights information. Several DAM vendors have extended into production and campaign workflow, arriving at the same territory from the governance side.

Commerce media and the retail networks need creative at volume, in many formats, for many placements, which is precisely the problem this category solves. A brand advertising across a dozen retail media networks is producing more creative variations than any traditional process supports.

The direction of travel is that creative production, asset governance, and media activation are converging into one workflow. The vendors in this Wave arrived at that convergence from the production side. DAM vendors arrive from governance. The media platforms arrive from activation, and they have the strongest distribution.

That three-way convergence is a reasonable explanation for why independence has been hard to maintain here.

What the sequence suggests

Three evaluations across six years, with a four-year gap between the second and third, describe a category that Forrester considered mature enough not to need annual coverage and important enough to return to.

The vendor churn suggests something more specific. Creative adtech has never been a stable market with defensible positions. It has been a useful capability that keeps getting absorbed by companies with better distribution, whether that is a media platform, a marketing cloud, or an advertising operations vendor.

For a buyer, the practical consequence is about contract length and exit terms rather than feature comparison. In a market where a third of the evaluated field disappears between editions, the questions worth asking are what happens to your templates, your variant history, and your performance data if the vendor is acquired, and how portable the creative logic you build actually is.

The technology in this category is genuinely good and getting better quickly. The companies providing it have been unusually fragile. Those two things have been true simultaneously for six years, and nothing in the current market structure suggests either is about to change.

Forrester does not endorse any vendor named here, and tier placement should not be read as a recommendation to buy.