Programmatic Creative Is Not a Tech Problem. It’s a Creative Problem.

Programmatic creative is the practice of automating the production and delivery of personalised ad creative at scale, matching the right message, format, and visual to the right audience at the right moment. Done well, it closes the gap between media precision and creative relevance. Done badly, it produces thousands of variations of the same mediocre idea, delivered with algorithmic efficiency to people who ignore all of them.

Most brands are doing it badly. Not because the technology is broken, but because they treat programmatic creative as a media problem rather than a creative one. The tools are sophisticated. The thinking behind the creative is not.

What Is Programmatic Creative and Why Does It Keep Underdelivering?

The promise of programmatic creative has always been compelling: serve the right creative to the right person at the right time, automatically, at a scale no human production team could match. In practice, most implementations fall well short of that promise. The creative is generic, the personalisation is surface-level, and the performance lift is modest at best.

I sat in a meeting a few years ago where a major technology vendor presented their AI-driven dynamic creative solution. The headline claim was a 90% reduction in cost per acquisition and a tripling of conversion rates. The room was impressed. I was sceptical. When I asked to see the baseline creative they had replaced, it was exactly what I expected: static, untested, visually cluttered, with a call to action buried in the corner. They had replaced genuinely poor creative with something marginally less poor, and the algorithm had done the rest. That is not a programmatic creative success story. That is a low baseline story. The AI did not create the uplift. Raising the floor did.

This distinction matters enormously. If your programmatic creative strategy begins with the technology and works backwards to the message, you will always be optimising the wrong variable.

How Programmatic Creative Actually Works

At its core, programmatic creative combines dynamic creative optimisation (DCO) with programmatic media buying. A DCO platform assembles ads in real time from a library of modular creative components: headlines, images, calls to action, offers, product feeds, and so on. The system then serves the combination most likely to perform for a given user, based on audience data, contextual signals, and historical performance.

The inputs that drive personalisation typically include audience segment data (demographics, behavioural, CRM), contextual signals (device, location, time of day, weather), and feed data (product inventory, pricing, availability). The better your inputs, the more meaningful the personalisation. But meaningful personalisation requires meaningful differences in your message, not just swapping a product image or changing a city name in the headline.

There is a useful framework for thinking about levels of personalisation in programmatic creative. At the most basic level, you are changing cosmetic elements: colours, images, product names. At the intermediate level, you are changing the message based on where someone is in the funnel or what category they have browsed. At the most sophisticated level, you are changing the entire creative rationale based on a deep understanding of what motivates different audience segments to act. Most brands operate at level one and call it personalisation.

If you are thinking about how programmatic creative fits into a broader go-to-market approach, the Go-To-Market and Growth Strategy hub covers the commercial frameworks that should be informing these decisions before you brief a single creative.

Why the Creative Strategy Has to Come Before the Technology

The biggest structural mistake I see in programmatic creative programmes is that they are set up by media teams and handed to creative teams as a production brief. The media team has built the audience architecture, defined the segments, and configured the DCO platform. The creative team is then asked to produce assets that fit the template. This is the wrong order of operations.

Creative strategy should define what needs to be said differently to different audiences, and why. The technology should then be configured to express those differences at scale. When you reverse this, you end up with a technically sophisticated system delivering the same message in slightly different wrappers. The algorithm optimises for the best-performing wrapper. You learn almost nothing about whether your message is working.

I spent time early in my career at a mid-sized agency where a client in financial services wanted to run a DCO campaign across their entire product range. The temptation was to build a massive matrix: twelve products, eight audience segments, four funnel stages. That is nearly 400 combinations before you account for format and placement. We stripped it back to three audience segments with genuinely different motivations, two funnel stages, and four products that actually drove volume. The creative team could write meaningfully different copy for those segments. The system had something real to work with. Performance was strong, and more importantly, we understood why.

Complexity is not sophistication. A programmatic creative programme with 400 variants and one underlying message is less effective than one with 20 variants built on three genuinely distinct propositions.

What Good Creative Architecture Looks Like in Practice

Building a programmatic creative programme that actually works requires thinking about creative architecture before you think about production volume. Creative architecture is the structured logic that determines what varies, what stays constant, and why.

Start with your audience segments and ask a simple question: does this segment have a meaningfully different reason to buy, a different objection to overcome, or a different relationship with the category? If the answer is yes, that segment warrants a different creative approach, not just a different image. If the answer is no, you do not need a separate segment in your DCO framework. You need better segmentation.

From there, define your creative constants: the elements that do not change regardless of audience or context. Brand identity, visual tone, and core value proposition typically sit here. Then define your creative variables: the elements that shift based on audience signal, funnel stage, or contextual trigger. Offers, proof points, calls to action, and product emphasis are good candidates. The ratio of constants to variables tells you how differentiated your creative strategy actually is.

One thing worth noting: feed-based personalisation, where the creative pulls in live product data, pricing, or inventory, is often the most straightforward form of programmatic creative to execute well. Retail and e-commerce brands have been doing this effectively for years. The challenge is not the technology; it is ensuring the product data is clean, the pricing logic is sound, and the creative frame around the feed is strong enough to carry the message. A beautifully formatted product card inside a weak creative concept is still a weak ad.

Dynamic Creative Optimisation Is a Testing Tool, Not a Creative Tool

DCO is frequently mischaracterised as a creative capability. It is not. It is a testing and optimisation mechanism. It tells you which combination of creative elements performs best against a defined metric. It does not tell you whether any of those elements are good. It does not tell you whether your message is resonating or your offer is compelling. It tells you which variant won in a specific context against a specific audience at a specific moment in time.

This distinction has real implications for how you interpret DCO results. A winning variant in a DCO test is not necessarily good creative. It is the best-performing option from the set you gave the system. If all your variants are weak, the winner is the least weak. I have seen brands use DCO results to validate creative decisions that should never have been made in the first place. The algorithm cannot save a bad brief.

The right way to use DCO is as a structured learning tool. Build variants that test a genuine hypothesis: does this audience respond better to a price-led message or a quality-led message? Does a product-first or lifestyle-first approach drive more consideration? Use the results to build creative intelligence over time, not just to find the least-bad version of your current campaign.

Vidyard’s research into go-to-market team performance touches on a related point: pipeline and revenue potential is often untapped not because of distribution gaps, but because of message gaps. The same is true in programmatic creative. The distribution infrastructure is rarely the constraint. The message is.

The Measurement Problem Nobody Wants to Talk About

Programmatic creative creates a measurement challenge that most teams handle poorly. When you have hundreds of creative variants running simultaneously across multiple audience segments and placements, attribution becomes genuinely difficult. Which element drove performance? Was it the headline, the image, the offer, the audience, the context, or some combination of all of them? DCO platforms will give you answers, but those answers are probabilistic, not definitive.

The temptation is to over-interpret the data. I have sat in performance reviews where teams have drawn very specific conclusions from DCO results that the data simply could not support. “Audiences in the 35-44 segment respond better to aspirational imagery” is a reasonable hypothesis. It is not a proven fact on the basis of one campaign cycle with a modest sample size. Treat DCO results as directional signals, not strategic mandates.

There is also a deeper measurement issue: programmatic creative optimises for measurable outcomes, typically clicks, conversions, or view-through events. These are not always the right outcomes to optimise for. If your campaign goal is brand consideration or category entry point ownership, optimising for click-through rate will pull your creative in the wrong direction. The system will learn to produce creative that gets clicked, which is not the same as creative that builds brand equity.

BCG’s work on go-to-market strategy and brand alignment makes a point that applies here: commercial performance and brand performance are not the same metric, and optimising for one at the expense of the other creates structural problems over time. Programmatic creative programmes that are measured purely on short-term conversion metrics tend to drift toward direct response execution even when the brief calls for something more brand-building.

Where Programmatic Creative Fits in the Broader Growth Strategy

Programmatic creative is a tactical execution capability. It is not a growth strategy. This sounds obvious, but a surprising number of brands treat it as if it were. They invest heavily in DCO platforms, creative production infrastructure, and audience data, and then wonder why the commercial impact is modest. The answer is usually that programmatic creative was deployed without a clear strategic rationale for what it was supposed to achieve.

The questions that should precede any programmatic creative investment are strategic, not technical. Who are you trying to reach and why? What do you want them to think, feel, or do differently as a result of seeing your creative? What is the role of paid media in your overall customer acquisition and retention model? How does creative personalisation serve your commercial objectives rather than just your media efficiency metrics?

Forrester’s framework for intelligent growth models is useful context here. Growth that is sustainable comes from building genuine competitive advantage, not from optimising the mechanics of existing channels. Programmatic creative can contribute to that, but only if it is expressing a genuinely differentiated proposition to a genuinely well-defined audience.

When I was running iProspect and growing the team from around 20 people to over 100, one of the things I noticed consistently was that the clients who got the most value from programmatic creative were the ones who had done the upstream strategic work. They knew their audience segments deeply, they had tested their messaging through other channels, and they came to programmatic creative with a clear point of view on what they wanted to say and to whom. The technology amplified a strategy that already existed. For clients who came to programmatic creative hoping it would generate the strategy, the results were consistently disappointing.

Market penetration strategy, as Semrush’s analysis of the topic outlines, requires clarity on which segments you are targeting and what your proposition is for each. Programmatic creative is one of the most powerful tools for executing against that clarity at scale. Without the clarity, it is an expensive way to produce a lot of noise.

If you want to think more rigorously about how programmatic creative connects to commercial strategy, the Go-To-Market and Growth Strategy hub is the place to start. The strategic frameworks there should be informing your creative decisions long before you brief a DCO platform.

What Separates the Programmes That Work From the Ones That Do Not

Having seen programmatic creative programmes across a wide range of categories and budget levels, the patterns that separate effective programmes from ineffective ones are consistent.

Effective programmes start with a creative brief, not a technology brief. The brief defines the audience insight, the message hierarchy, the creative rationale, and the success criteria. The technology is configured to express that brief at scale. Ineffective programmes start with the platform capabilities and ask creative teams to fill the slots.

Effective programmes have a clear taxonomy of what varies and why. Every variable in the creative matrix exists because there is a strategic reason for it to exist. Ineffective programmes have variables because the platform supports them, not because the strategy requires them.

Effective programmes treat DCO results as learning, not validation. They use performance data to refine their understanding of what messages resonate with which audiences, and they feed that learning back into the creative strategy. Ineffective programmes use DCO results to confirm decisions that were already made.

Effective programmes have a clear view of what they are optimising for and why that metric connects to a business outcome. Ineffective programmes optimise for whatever the platform measures by default.

Forrester’s thinking on agile scaling is relevant here: the discipline that makes scaling work is the same discipline that makes programmatic creative work. Clear principles, structured decision-making, and a willingness to stop doing things that are not working.

One more thing worth saying plainly: programmatic creative is not a cost reduction tool. It can improve media efficiency over time, but the upfront investment in creative strategy, production infrastructure, data integration, and platform management is significant. Brands that approach it primarily as a way to reduce creative production costs tend to cut corners on the strategic work that makes the whole thing function. The economics only work if the programme is performing, and the programme only performs if the creative strategy is sound.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions