Programmatic Marketing Is Buying Media. It Is Not a Strategy

Programmatic marketing is the automated buying and selling of digital advertising inventory, using data and algorithms to serve ads to specific audiences in real time. It replaced the manual insertion order process that defined media buying for decades, and it now accounts for the vast majority of digital display spend globally. But understanding what programmatic is, and knowing how to use it well, are two very different things.

Most brands are running programmatic. Far fewer are running it with any real strategic discipline. The technology is sophisticated. The thinking behind how most organisations deploy it often is not.

What Programmatic Marketing Actually Does

At its core, programmatic automates the process of buying ad space. A publisher makes inventory available through a supply-side platform. An advertiser bids for that inventory through a demand-side platform. The whole transaction, including the auction, the decision, and the ad serving, happens in milliseconds. The promise was efficiency, transparency, and better targeting. The reality has been more complicated.

When programmatic started scaling in the early 2010s, the pitch was compelling. You could reach the right person, at the right time, on the right device, at a fraction of the cost of traditional media. Agencies and clients bought in hard. I was running teams during that period, and I watched the industry pour budget into programmatic display with enormous confidence and relatively little scrutiny. The CPMs were low. The dashboards looked good. The attribution models pointed in the right direction. Everyone was happy.

Then clients started asking harder questions. Why was brand recall not improving? Why were new customer acquisition rates flat even as retargeting numbers looked strong? Why did some campaigns produce impressive click-through rates but no measurable revenue uplift? The technology was working. The strategy behind it was not.

The Retargeting Trap Most Brands Still Have Not Escaped

Retargeting is the most widely used programmatic tactic. It is also the most overvalued. The logic is straightforward: someone visited your site, showed interest, did not convert, so you follow them around the internet with ads until they come back and buy. It feels like smart marketing. In many cases, it is not.

The problem is attribution. When someone who has already visited your site converts after seeing a retargeting ad, the platform credits the ad. But that person was already in your funnel. They already knew who you were. A meaningful proportion of them would have returned and converted anyway. You are paying to take credit for something that was going to happen regardless.

Earlier in my career I was guilty of overvaluing exactly this kind of lower-funnel activity. It looked like performance. The numbers were clean. But when we ran holdout tests on some of our retargeting campaigns, the incremental lift was often far smaller than the attributed revenue suggested. We were not generating new demand. We were just standing at the door of a shop that customers were already walking into, charging ourselves for the privilege of opening it.

That does not mean retargeting has no value. It does. But it should be sized accordingly, and it should never be the dominant programmatic investment for a brand that needs to grow its customer base. Growth requires reaching people who do not yet know you exist, not just converting people who have already found you.

If you want a broader view of how programmatic fits within a growth framework, the Go-To-Market and Growth Strategy hub covers the strategic context that media decisions need to sit inside.

Prospecting at Scale: Where Programmatic Can Actually Drive Growth

The more interesting and more difficult use of programmatic is prospecting. Reaching new audiences. Building demand upstream, before someone has shown any intent. This is where the technology has genuine potential to contribute to growth, and where most brands underinvest relative to their retargeting and lower-funnel activity.

Effective prospecting through programmatic requires three things that are harder than they sound: a clear definition of who you are trying to reach, creative that is built for awareness rather than conversion, and measurement frameworks that do not penalise activity that does not produce an immediate click.

The audience definition piece is where I see the most consistent failure. Brands either go too broad, buying cheap inventory against loosely defined demographic segments, or they rely too heavily on third-party data that is of questionable quality. The deprecation of third-party cookies has accelerated this problem. Audience signals that programmatic platforms were built around are eroding, and many brands have not yet built the first-party data infrastructure to replace them.

When I was scaling a performance agency from around 20 people to over 100, one of the hardest conversations we had with clients was around audience strategy. Everyone wanted to know their cost per acquisition. Far fewer wanted to invest in understanding who their best customers actually were, where they spent time online, what they cared about, and how to reach people who looked like them before they had ever searched for the brand. That upstream thinking is what separates programmatic that builds a business from programmatic that just processes existing demand.

The mechanics of market penetration are worth understanding here. Programmatic prospecting is one of the few digital channels that can genuinely expand your addressable market if it is set up correctly. Most brands are not using it that way.

The Structural Problems That Undermine Most Programmatic Campaigns

Programmatic has three structural problems that have been well documented but are still routinely underestimated in practice: ad fraud, brand safety, and viewability. None of them are new. All of them continue to drain budgets that brands believe are being spent on real audiences.

Ad fraud is the practice of generating fake impressions, clicks, or conversions to extract money from advertisers. The scale of it has been consistently underestimated. Sophisticated fraud operations use bot networks to simulate real user behaviour, including visiting websites, spending time on pages, and interacting with ads. The metrics look legitimate. The humans do not exist.

Brand safety refers to the risk of your ads appearing alongside content that is harmful, offensive, or simply inappropriate for your brand. Programmatic buys inventory at scale and at speed. Without the right controls, your ad can end up next to content you would never consciously choose to be associated with. I have seen this happen to clients across industries. The reputational cost is real, and the damage is often disproportionate to the amount of money involved.

Viewability is the question of whether an ad was actually seen. An impression is counted when an ad is served, not when it is viewed. Ads that load below the fold, in tabs that are never opened, or in positions where they are immediately scrolled past, all generate impressions that register in your reporting but have zero chance of influencing anyone. Industry standards define a viewable impression as one where at least 50% of the ad is in view for at least one second. That is a very low bar, and a significant proportion of programmatic inventory does not even clear it.

Managing these three risks requires active oversight. You need inclusion lists rather than exclusion lists where possible. You need third-party verification. You need to be willing to accept lower reach in exchange for higher quality inventory. The platforms will not optimise for this by default, because their incentive is to spend your budget, not to protect it.

The Supply Chain Problem Nobody Talks About Enough

There is a money flow problem at the heart of programmatic that the industry has been slow to address honestly. When a brand spends a pound on programmatic display, a significant portion of that pound does not reach the publisher. It is absorbed by the layers of technology and intermediaries that sit between the advertiser and the inventory: the DSP, the SSP, the data providers, the verification tools, the ad servers. The exact proportion varies, but the waste is structural and substantial.

This matters for two reasons. First, it means the effective CPM you are paying for publisher inventory is higher than it appears in your reporting. Second, it means publishers receive less revenue per impression than the advertiser believes they are paying, which creates incentives for publishers to accept lower-quality inventory or to fill their ad slots with content that maximises yield rather than audience quality.

The response to this from sophisticated buyers has been a shift toward supply path optimisation: deliberately reducing the number of intermediaries in the chain, building direct relationships with premium publishers, and using private marketplaces rather than open exchange inventory. This is the right direction. It requires more effort than simply letting a DSP run on autopilot, but it produces meaningfully better outcomes.

I spent time working across financial services clients where compliance and brand reputation were non-negotiable. The open exchange was simply not an appropriate environment for their advertising. Private marketplace deals with verified premium publishers were the only viable route. The CPMs were higher. The waste was lower. The outcomes were better. That trade-off is often worth making, even for brands that do not have the same regulatory constraints.

Creative Is Still the Biggest Lever Nobody Optimises

The industry conversation around programmatic is almost entirely focused on targeting, bidding, and measurement. Creative gets a fraction of the attention, despite being the single biggest driver of whether an ad actually works. You can have perfect audience targeting and flawless brand safety controls, but if the ad itself is forgettable, it will not move anything.

Dynamic creative optimisation is the technology that is supposed to address this. DCO allows you to serve different versions of an ad, different headlines, images, calls to action, to different audience segments, and to optimise toward the combinations that perform best. The technology is genuinely useful. The problem is that most brands using DCO are optimising around a very narrow set of creative variables, and they are optimising for clicks rather than for outcomes that matter.

The deeper issue is that programmatic display is an interruption medium. People are not there to see your ad. They are there for the content around it. The creative has to earn attention in a context where attention is not on offer. Most programmatic creative does not do this. It is built for conversion at the bottom of the funnel, not for building memory structures at the top. It is transactional when it should be memorable.

When I judged at the Effie Awards, the campaigns that stood out in the digital and performance categories were almost always the ones where the creative team and the media team had worked together from the beginning. The creative was built for the context, not retrofitted to it. The targeting was informed by what the creative could credibly say to different audience segments. That integration is rare. It should not be.

Measurement: Honest Approximation Over False Precision

Programmatic generates an enormous amount of data. Impressions, clicks, view-through conversions, post-click conversions, frequency, reach, CPM, CPC, ROAS. The dashboards are full. The understanding is often thin.

The measurement problem in programmatic has two dimensions. The first is attribution: which touchpoints actually caused a conversion, and how much credit should each one receive. The second is incrementality: how much of what you are measuring would have happened anyway without the advertising.

Last-click attribution, which is still the default in many organisations, systematically overstates the contribution of lower-funnel programmatic activity and understates the contribution of upper-funnel brand activity. It is a measurement choice that shapes budget allocation, and it consistently pushes investment toward channels that look efficient on paper but are largely harvesting demand that other activity created.

Multi-touch attribution models are an improvement but they are not a solution. They are still operating on modelled assumptions about how touchpoints interact. Incrementality testing, through holdout groups and geo-based experiments, is the closest thing to honest measurement that most brands can realistically access. It is harder to set up and harder to explain to a CFO than a clean ROAS number, but it tells you something closer to the truth.

The principle I try to apply is honest approximation over false precision. A measurement framework that acknowledges its limitations but points in the right direction is more useful than one that produces confident numbers built on questionable assumptions. Programmatic platforms are not neutral reporters of what happened. They are interested parties in how performance is attributed. That context matters.

Understanding how to structure feedback loops into your measurement approach is worth exploring. Growth loop thinking offers a useful frame for connecting media activity to the broader signals that indicate whether a business is actually growing.

How to Build a Programmatic Setup That Actually Works

Most programmatic setups are built around platform defaults. The DSP recommends a bidding strategy. The account manager suggests an audience segment. The campaign goes live. The dashboard fills up with numbers. Nobody stops to ask whether any of this connects to a business objective.

A programmatic setup that works starts with a clear brief: who are you trying to reach, what do you want them to think or do, and how will you know if it worked. That sounds obvious. It is consistently skipped. The technology is so easy to activate that the strategic thinking that should precede it often gets compressed or bypassed entirely.

From there, the decisions follow in a logical sequence. Audience definition based on first-party data and validated third-party signals. Inventory strategy that balances reach with quality, using private marketplaces for premium environments and open exchange with strong controls for broader reach. Creative that is built for the specific context and audience, not repurposed from another channel. Measurement that includes both in-flight optimisation metrics and longer-term outcome measures. And governance: who is checking the brand safety controls, who is reviewing the placement reports, who is responsible for the quality of what is being bought.

That last point matters more than most organisations acknowledge. Programmatic runs at scale and at speed. Problems compound quickly. A brand safety incident that could have been caught by a weekly placement review can generate significant reputational damage if it runs unchecked for a month. The automation does not remove the need for human oversight. It changes what that oversight needs to look like.

Agile ways of working have a role here. Forrester’s work on agile scaling is relevant to how teams structure their review and optimisation cadences. The principle of frequent, small adjustments based on real data applies directly to programmatic campaign management.

Where Programmatic Fits in a Broader Go-To-Market Plan

Programmatic is a channel. It is a capable and scalable channel, but it does not make strategy. It executes strategy. The distinction matters because a lot of marketing organisations treat programmatic as if it is self-directing: set the objectives in the platform, let the algorithm optimise, read the results. That is not strategy. That is delegation to a machine that has different incentives from yours.

In a well-constructed go-to-market plan, programmatic has a specific role. It can build awareness at scale in new markets or with new audience segments. It can support a product launch by reaching defined prospect pools with sequential messaging. It can reinforce brand positioning across the funnel. It can retarget with relevant creative at the right frequency. None of those things happen by default. They happen because someone made deliberate choices about what programmatic is supposed to do and built the campaign architecture accordingly.

The brands that use programmatic well tend to be the ones that treat it as one component in a connected system, not as a standalone performance channel. They think about how programmatic activity connects to paid search, to organic, to content, to CRM. They think about the customer experience across channels, not just the performance of each channel in isolation. That systems thinking is what separates programmatic that contributes to growth from programmatic that just generates impressions.

Go-to-market strategy is the broader frame that media decisions need to sit inside. If you are working through how programmatic connects to your wider growth approach, the Go-To-Market and Growth Strategy hub covers the strategic foundations that make channel decisions coherent rather than reactive.

For brands in growth mode, the documented examples of growth-stage marketing are worth reviewing for context on how channel mix decisions tend to evolve as businesses scale. The pattern is consistent: early-stage businesses over-index on lower-funnel performance, then discover that growth stalls without upstream investment. Programmatic prospecting is one of the mechanisms that can address that stall, if it is deployed with the right brief.

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