Programmatic Advertising Strategy: Stop Buying Audiences You Already Own

Programmatic advertising strategy is the structured approach to planning, buying, and optimising automated media across channels, with the goal of reaching the right audience at the right cost, at scale. Done well, it connects real business objectives to media decisions. Done poorly, it burns budget retargeting people who were already going to convert.

Most programmatic programmes sit somewhere in the middle: technically functional, commercially underwhelming. The problem is rarely the technology. It is the strategy behind it.

Why Most Programmatic Strategies Are Built Backwards

When I was running an agency and we onboarded a new client with a sizeable programmatic budget, the first question I always asked was: who are you trying to reach that you are not reaching today? The answer was almost always a variation of “we want to reach people who are interested in buying our product.” Which is fine, but that is not a strategy. That is a targeting brief.

The distinction matters because programmatic platforms are very good at finding people who look like your existing customers. They are considerably less good, without deliberate instruction, at finding people who have never considered you but should. The algorithm optimises for what you reward it for. If you reward it for conversion signals, it will find converters. It will also find people who were going to convert anyway.

This is the lower-funnel trap. Earlier in my career I overvalued performance metrics precisely because they looked so clean. Cost per acquisition, return on ad spend, conversion rate: all tidy, all attributable, all pointing in the right direction. What they were not telling me was how much of that activity was incremental. A lot of it was not. The people converting had already decided. We were just the last ad they saw before they clicked.

Think about a clothes shop. Someone who has already tried something on is far more likely to buy than someone who has never walked through the door. Retargeting the trial group relentlessly makes the numbers look good. But if you never bring new people through the door, you are managing a shrinking pool. Programmatic strategy has to account for both.

If you are working through how programmatic fits into a broader commercial plan, the Go-To-Market and Growth Strategy hub covers the full picture, from market penetration to channel sequencing.

What Audience Strategy Actually Means in Programmatic

Audience strategy in programmatic is not the same as audience targeting. Targeting is the execution layer. Strategy is the decision about which audiences to prioritise, in what order, with what message, and why.

A useful way to think about this is in three tiers:

Tier one: known audiences. These are people you have a direct data relationship with. CRM lists, site visitors, app users, past purchasers. This is your highest-confidence targeting, and your retargeting pool. The risk here is over-serving. Most brands retarget too heavily and for too long. Someone who visited your pricing page six weeks ago and did not convert has almost certainly made their decision. Continuing to serve them ads is not persuasion, it is noise.

Tier two: modelled audiences. Lookalike and contextual segments built from your first-party data or from publisher and platform signals. This is where programmatic earns its keep in growth terms, because you are reaching people who match the profile of a buyer but have not yet engaged. The quality of the seed data matters enormously here. If your CRM is full of low-value customers, your lookalike will find more of them.

Tier three: net new audiences. Broad contextual targeting, interest-based segments, or publisher-specific buys designed to reach people outside your current consideration set. This is where brand-building happens in programmatic. It is also where most performance-focused teams underinvest, because the signal-to-noise ratio is harder to read. Understanding how market penetration works makes clear why this tier cannot be ignored if growth is the actual objective.

The strategic error most brands make is treating tier one as the whole programme. It is the easiest to measure and the easiest to justify in a performance review. It is also the least likely to drive growth.

Supply Path and Brand Safety: The Unglamorous Decisions That Actually Matter

When I judged the Effie Awards, one thing that struck me about the campaigns that performed best commercially was how deliberate they were about context. Not just what they said, but where they showed up. That deliberateness rarely comes from a DSP’s default settings.

Supply path optimisation is one of the least discussed and most commercially significant decisions in programmatic strategy. The open exchange is cheap. It is also opaque, and a meaningful proportion of inventory on the open exchange is not what it claims to be. Impression fraud, domain spoofing, and made-for-advertising sites are not edge cases. They are structural features of an unregulated supply chain.

The practical implication is that your CPM on the open exchange may look efficient while delivering impressions that no human ever saw. That is not a measurement problem. It is a supply problem. The answer is not to abandon programmatic buying, but to be deliberate about where you buy.

Private marketplace deals and programmatic guaranteed arrangements give you publisher-level transparency and often better audience quality, at a higher CPM. For most brands, the effective CPM after accounting for invalid traffic on the open exchange is not as different as it looks. The premium for quality supply is frequently worth paying.

Brand safety is a related but distinct issue. Most DSPs offer keyword blocklists and content category exclusions. These are blunt instruments. A news site that covers crime and politics is not unsafe for most advertisers, but a keyword-level block will exclude it. The result is that brand safety tools often push spend toward lower-quality inventory while giving the impression of protection. Review your inclusion lists as carefully as your exclusion lists.

Frequency Management: The Problem Nobody Wants to Talk About

Frequency capping is one of those settings that gets configured once during campaign setup and then forgotten. It should not be. Uncapped or poorly managed frequency is one of the most reliable ways to waste programmatic budget, and one of the fastest ways to damage how people feel about your brand.

The evidence from brand tracking studies and attention research consistently points in the same direction: there is an optimal exposure range for any given creative execution, and beyond it, sentiment deteriorates. The ad that felt relevant at impression three feels intrusive at impression twelve. The person who might have converted starts actively avoiding your brand.

The challenge in programmatic is that frequency is hard to control across channels. You can cap frequency within a DSP, but if the same person is also seeing your ads through a social platform, a publisher direct buy, and a video network, your effective frequency across their experience is much higher than any single platform reports. Cross-channel frequency management requires a unified view, which is technically difficult and commercially underinvested.

A practical starting point: set conservative frequency caps within each platform, review reach and frequency curves weekly in the early stages of a campaign, and treat any creative that is running at high frequency for more than two weeks as a candidate for rotation. The creative refresh cadence in programmatic should be faster than most teams plan for.

Measurement Frameworks That Do Not Lie to You

Programmatic measurement is where the most comfortable fictions live. Last-click attribution in a programmatic context is almost meaningless, because the channel is rarely the last touchpoint in any meaningful sense. View-through attribution, which credits a conversion to a display impression the user may have seen for 0.3 seconds three weeks ago, is even more suspect.

I have sat in enough reporting reviews to know that most programmatic performance dashboards are telling you what the model says happened, not what actually happened. The model reflects the attribution logic you chose, not the market. When I was managing large media accounts, we ran the same campaign with three different attribution windows and got three completely different ROAS figures. All three were technically accurate. None of them was the truth.

Honest programmatic measurement requires a few things:

Incrementality testing. The only real answer to “did this programmatic activity cause conversions” is to run a holdout test and compare. Ghost bidding, geographic holdouts, and matched market tests are all valid approaches. They are more work than pulling a platform report, and they will almost certainly show lower returns than your attribution model suggests. That is the point. You need to know what you are actually buying.

Reach metrics alongside conversion metrics. If your programmatic programme is genuinely reaching new audiences, you should see reach growing among people who do not appear in your CRM. If your reach metrics are flat and your conversion metrics look healthy, you are probably just retargeting your way through an existing pool.

Brand health tracking. For any programme investing meaningfully in upper-funnel programmatic, brand awareness, consideration, and purchase intent metrics should be tracked alongside performance metrics. These move slowly, but they are the leading indicators of whether your audience strategy is working. Platforms like Hotjar can help surface behavioural signals that complement what your programmatic platform reports.

Marketing does not need perfect measurement. It needs honest approximation. The goal is to be less wrong over time, not to manufacture precision that the data cannot support.

Creative Strategy in Programmatic: The Variable That Gets Ignored

Programmatic conversations are dominated by targeting, bidding, and measurement. Creative is treated as an input, not a strategic variable. This is a significant mistake.

The targeting determines who sees your ad. The creative determines what they do with it. A technically perfect programmatic setup with weak creative will underperform a less sophisticated setup with strong creative, consistently. I have seen this play out enough times that it is no longer surprising, but it still does not get the attention it deserves in most programmatic planning conversations.

Dynamic creative optimisation gets a lot of attention as a solution here, and it can be useful. But DCO is a testing and personalisation tool, not a substitute for a clear creative idea. Algorithmically assembling combinations of headlines and images will not rescue a campaign that has nothing interesting to say.

The more productive framing is to treat creative as a campaign variable with its own testing plan. What message is right for tier-one retargeting audiences versus tier-three net new audiences? What format works for awareness versus consideration? How quickly does creative fatigue set in for each segment? These are strategic questions, not production questions, and they should be answered before the campaign launches.

Creator-led content is increasingly relevant in programmatic contexts. Brands running creator content through paid amplification, including programmatic channels, often see stronger engagement signals than standard display formats. Later’s research on creator-led go-to-market campaigns is worth reviewing if you are considering how to blend creator content with programmatic distribution.

How to Structure a Programmatic Strategy That Serves Growth

A programmatic strategy that actually serves growth has to be built around a commercial objective, not a channel objective. “Improve ROAS” is a channel objective. “Grow market share among 25-40 year old first-time buyers in three target markets” is a commercial objective. The difference determines everything: which audiences you prioritise, how you allocate budget across funnel stages, what measurement framework you use, and what success looks like at the end of a quarter.

The practical structure I have used across different categories looks like this:

Start with the commercial brief. What is the business trying to achieve? What does the customer acquisition or growth problem actually look like? This is the foundation. Without it, every programmatic decision is arbitrary.

Define the audience architecture. Map out the three tiers described earlier. Assign budget proportions based on where the growth opportunity actually is, not where the measurement is easiest. If you are trying to grow, tier three deserves more budget than most performance-led organisations give it.

Choose supply paths deliberately. Decide which inventory environments are appropriate for the brand and the campaign objective. Set inclusion lists, not just exclusion lists. Negotiate PMP deals with publishers whose audiences match your tier-two and tier-three targets.

Build a measurement plan before you launch. Decide how you will evaluate incrementality. Set frequency caps. Agree on what the reporting cadence will be and what decisions each report is designed to inform.

Plan creative rotation from the start. Know when you will refresh creative and what signals will trigger that decision. Do not wait until performance drops to address creative fatigue.

Scaling this kind of approach requires organisational discipline as much as technical capability. BCG’s work on scaling agile practices is relevant here: the challenge of moving from a technically functional programmatic setup to one that genuinely drives growth is as much about how teams make decisions as it is about the platforms they use.

If you are looking at how programmatic fits within a wider growth framework, the articles across the Go-To-Market and Growth Strategy hub cover channel sequencing, market penetration, and commercial planning in more depth.

The Uncomfortable Truth About Programmatic Efficiency

Programmatic advertising is genuinely efficient at doing what you tell it to do. The problem is that what most organisations tell it to do is optimise for signals that are easier to measure than they are meaningful. Conversion-optimised campaigns find converters. They do not necessarily find growth.

The most commercially effective programmatic programmes I have been involved with were not the ones with the best ROAS figures. They were the ones where someone had made a deliberate decision about how much of the budget was being used to capture existing demand versus create new demand, and had built a measurement framework honest enough to tell the difference.

That requires resisting the pull of the dashboard. It requires making the case internally for investment that will not show clean returns in the first quarter. And it requires being willing to say, clearly, that a lot of what programmatic is credited for in most attribution models was going to happen anyway.

None of that is comfortable. But it is the work. And it is what separates a programmatic strategy from a programmatic budget.

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