Programmatic Advertising Is Buying Media. It Is Not a Strategy

Programmatic advertising is the automated buying and selling of digital ad inventory through real-time bidding systems, demand-side platforms, and data-driven audience targeting. It replaced the manual insertion order process that defined display advertising for its first decade, and it now accounts for the vast majority of digital display spend globally. But knowing what it is and knowing how to use it well are two very different things.

Most brands treat programmatic as a performance channel. They optimise toward last-click conversions, bid on retargeting audiences, and measure success in cost-per-acquisition. That framing misses most of what programmatic can actually do, and it systematically undervalues the role of reach, frequency, and brand exposure in driving the outcomes performance dashboards later claim credit for.

What Programmatic Advertising Actually Is

At its core, programmatic advertising is a technology stack that connects advertisers to publishers through automated auctions. When a user loads a webpage or opens an app, an auction happens in milliseconds. Advertisers bid for that impression based on who the user is, what context they are in, and what the advertiser is willing to pay. The highest bidder wins the impression. The ad serves. The user sees it, or doesn’t.

The main components are the demand-side platform (DSP), which is where advertisers manage their buying; the supply-side platform (SSP), which is where publishers manage their inventory; and the ad exchange, which connects the two. Data management platforms (DMPs) and, increasingly, clean room environments layer audience intelligence on top of this infrastructure. Add in verification tools, viewability measurement, and brand safety filtering, and you have a reasonably complex ecosystem even before you write a single line of creative.

The formats available through programmatic have expanded significantly. Display banners were the original use case. Now programmatic covers video pre-roll and mid-roll, connected TV, digital out-of-home, audio, native, and even some digital print environments. The underlying auction mechanics are similar across formats, but the execution considerations, creative requirements, and measurement approaches differ considerably.

Programmatic sits within the broader discipline of go-to-market execution. If you are thinking about where it fits within a full growth strategy, the wider context matters. The Go-To-Market and Growth Strategy hub covers how paid media channels, including programmatic, connect to commercial objectives across different business stages.

Why the Performance Framing Gets Programmatic Wrong

I spent the early part of my career deeply invested in lower-funnel performance metrics. Conversion rates, cost per lead, return on ad spend. It felt rigorous. It felt accountable. And in a lot of ways, it was. But over time I started noticing a pattern that took a while to name properly: a significant portion of what performance channels were being credited for was going to happen anyway.

Think about a clothes shop. Someone who walks in and tries something on is far more likely to buy than someone who just browses the window. Performance marketing, particularly retargeting, is very good at finding the people who have already tried something on. It puts an ad in front of someone who visited your site, searched your brand name, or compared your product against a competitor. They were already warm. The conversion was likely. The attribution model assigned the credit to the last click, and the performance channel looked brilliant.

What that framing misses is the question of how those people got warm in the first place. Programmatic, used well, is one of the most powerful tools for creating that warmth at scale. It can reach genuinely new audiences, build frequency against high-value segments, and shift brand perception in ways that make every downstream channel more effective. But only if you use it that way.

The problem is that most programmatic setups are optimised for the last step rather than the experience. Retargeting pools get prioritised because they convert. Prospecting campaigns get cut because they don’t convert immediately. The budget migrates toward the bottom of the funnel, the top dries up, and six months later the retargeting pool shrinks because there are fewer new people entering the consideration set. At that point, everyone wonders why performance has declined.

This is not a technology problem. It is a measurement and incentive problem. Programmatic platforms will optimise toward whatever signal you give them. If you give them conversion events, they will find converters. If those converters were already going to convert, the platform looks efficient and the business is treading water.

The Full-Funnel Case for Programmatic

Programmatic’s real advantage is precision at scale across every stage of the buying cycle, not just the bottom layer. That means thinking about it differently depending on the objective.

At the awareness stage, programmatic can reach defined audience segments across thousands of publishers simultaneously. You can target by demographic, interest, behavioural signals, contextual environment, or a combination. Connected TV and digital audio have made this even more powerful, because you can reach audiences in lean-back environments where attention is higher and clutter is lower. This is not direct response. You are not expecting someone to click. You are building familiarity, and familiarity is the precondition for consideration.

At the consideration stage, programmatic can serve more detailed creative to people who have shown some signal of interest, whether that is a category search, a competitor visit, or engagement with relevant content. Sequential messaging, where you serve different creative in a planned order based on where someone is in the funnel, is one of the more underused capabilities available in most DSPs. Most advertisers set up one audience and one creative and call it done.

At the conversion stage, retargeting and high-intent audience targeting make sense. But this should be the smallest slice of the budget, not the largest. It should also carry the most scrutiny. Retargeting pools contain a lot of people who were never genuinely in-market. Frequency caps matter. Creative fatigue is real. Serving the same banner to someone forty times does not increase purchase intent; it increases annoyance.

The brands that get the most from programmatic are the ones that think about the full funnel as a connected system, not a collection of independent campaigns. Each stage feeds the next. Awareness builds the pool. Consideration qualifies it. Conversion captures it. If you only fund the last stage, you are fishing in a pond you stopped stocking.

Audience Strategy Is Where Programmatic Wins or Loses

Technology does not replace thinking. It amplifies it. A well-defined audience strategy in programmatic will outperform a poorly defined one regardless of which DSP you use, how sophisticated your bidding algorithm is, or how much you spend on data.

The first question is always: who are you actually trying to reach? Not in a vague demographic sense, but in a commercially meaningful one. Who is most likely to buy? Who is most likely to become a high-value customer? Who is currently buying from a competitor and could be switched? These questions should shape audience construction, not be answered after the fact by whatever the platform’s default segments happen to be.

First-party data is the most valuable input you have. CRM lists, customer purchase data, email engagement data, site behaviour data. These tell you who your real customers are, and they give you a foundation for lookalike modelling that is grounded in actual commercial performance rather than platform-defined proxies. As third-party cookies have declined and privacy regulations have tightened, the gap between advertisers with strong first-party data infrastructure and those without has widened considerably.

Contextual targeting has come back into focus as a result. Serving ads in environments that are relevant to the product or service, rather than following a user around based on their behavioural profile, is both more privacy-compliant and, in many cases, more effective than it was given credit for. Context signals intent. Someone reading a detailed review of running shoes is probably thinking about running shoes. That is a useful signal even without a cookie.

The combination of first-party data, contextual signals, and platform audience tools gives most advertisers more targeting capability than they know what to do with. The discipline is in being selective. Narrow, well-defined audiences with relevant creative will outperform broad audiences with generic creative almost every time. The temptation to reach everyone is usually a sign that the audience strategy has not been thought through.

Brand Safety and Inventory Quality Are Not Optional

One of the things I have had to explain to more clients than I would like is that cheap impressions are often cheap for a reason. Programmatic’s open exchange contains a significant volume of low-quality inventory: sites with minimal human traffic, environments that no brand would choose if they saw them in advance, and ad placements that technically serve but are never actually seen by a real person.

Viewability is the most basic quality filter. An impression that is never in view cannot do anything. Viewability standards vary by format, but a display ad that loads below the fold and is never scrolled into view has zero value. Most DSPs report viewability metrics, but the default settings often prioritise volume over quality. You have to actively configure campaigns to weight toward viewable inventory.

Brand safety goes further. This is about where your ads appear, not just whether they are seen. Appearing alongside extremist content, misinformation, or low-quality clickbait does not just waste money. It associates your brand with that environment. Most advertisers use keyword blocklists and category exclusions, but these are blunt instruments. Private marketplaces and curated deals, where you buy inventory directly from vetted publishers through programmatic pipes, give you much more control.

Invalid traffic is the third layer. Ad fraud, bot traffic, and click farms inflate impression and click numbers without delivering any real audience. Verification tools from companies like IAS and DoubleVerify add a layer of protection, but they are not perfect. Periodic auditing of where your spend is actually going, at the domain and app level, is the only way to catch what automated filters miss.

None of this is glamorous. It does not make for an exciting agency presentation. But ignoring it means a meaningful proportion of your programmatic budget is doing nothing, or worse, actively damaging brand perception. I have seen audit reports that revealed a significant share of spend going to domains that no sensible marketer would have chosen consciously. The autopilot had been running unchecked.

The In-Housing Question

The trend toward in-housing programmatic capability has been real and, in some cases, well-founded. There are genuine advantages to bringing the function inside: better data access, faster iteration, lower fees, and greater transparency into where money is actually going. For large advertisers with significant programmatic budgets, the economics can work.

But the in-housing conversation often underestimates what it actually takes. Running a DSP well requires specialist expertise that is genuinely hard to hire and retain. The platforms are complex, the auction dynamics change constantly, and the data infrastructure needed to support sophisticated audience strategies is not trivial to build. A lot of brands have in-housed programmatic buying only to find that the person operating the platform is less experienced than the agency team they replaced, and the cost savings are offset by performance decline.

The middle path that tends to work best is a hybrid model: in-house strategic control and data ownership, with specialist execution support where the technical complexity is highest. This keeps the brand close to the data, maintains transparency, and still accesses the depth of expertise that full-time in-house teams often struggle to sustain.

Transparency is the non-negotiable regardless of model. You should know your media costs separately from your technology costs and your agency fees. You should know what markup is applied to inventory. You should have access to impression-level data if you want it. If any part of that picture is opaque, that is a problem worth solving before you optimise anything else.

Measurement: What Programmatic Can and Cannot Tell You

Programmatic generates an enormous volume of data. Impressions, clicks, viewability rates, completion rates for video, frequency, reach, cost per thousand, cost per click, cost per action. The dashboards are detailed and they update in near real-time. This creates a false sense of certainty that I think is one of the more dangerous aspects of the channel.

The data tells you what happened within the platform. It does not tell you why it happened, whether it would have happened anyway, or what the counterfactual looks like. Last-click attribution, which is still the default in many setups, systematically overvalues the final touchpoint and undervalues everything that came before it. View-through attribution, where a conversion is credited to an ad that was served but not clicked, overcorrects in the other direction and inflates programmatic’s apparent contribution.

The honest answer is that measuring programmatic’s true contribution to business outcomes is genuinely difficult. Incrementality testing, where you run controlled experiments with holdout groups to measure the difference between exposed and unexposed audiences, is the most rigorous approach available. It is also more expensive and slower than reading a dashboard. Most advertisers do not do it consistently, which means they are making budget decisions based on attribution models that have known flaws.

That does not mean measurement is hopeless. It means you need to be honest about what each metric is actually telling you. Viewability and completion rates tell you about ad delivery quality. Click-through rates tell you about creative relevance and audience match, though they are a weak proxy for business impact. Brand lift studies tell you about awareness and perception shifts. Sales lift studies and geo-based tests tell you about revenue impact. Using a mix of these, rather than relying on a single dashboard metric, gives a more honest picture.

I judged the Effie Awards for several years, and one thing that became clear quickly is that the campaigns with the most convincing effectiveness cases were the ones that used multiple measurement approaches rather than a single metric. They triangulated. They were honest about what they could and could not prove. That rigour is exactly what most programmatic measurement setups are missing.

Creative Is the Variable Most Programmatic Setups Ignore

There is a tendency to treat programmatic as a media problem: get the targeting right, get the bidding right, get the inventory quality right, and the results will follow. That framing ignores the variable that often has the largest impact on performance: the creative.

A well-targeted ad with weak creative is still a weak ad. It reaches the right person and fails to do anything with the opportunity. In display advertising particularly, where attention is low and creative real estate is limited, the quality of the message and the clarity of the value proposition matter enormously. The banner that tries to say five things says nothing. The one that makes a single, specific, relevant point at least has a chance.

Dynamic creative optimisation (DCO) is the programmatic answer to creative testing at scale. It allows you to serve different combinations of headline, image, and call to action to different audience segments, and to optimise toward the combinations that perform best. Used well, it is a genuinely powerful tool. Used lazily, it becomes an excuse not to think about creative strategy, because the algorithm will figure it out. It will not. The algorithm optimises within the options you give it. If all the options are mediocre, it will find the least mediocre one.

The creative briefing process for programmatic also tends to be underdeveloped. Creative teams are often briefed on the format requirements (sizes, file weights, animation length) without being given clear audience context, funnel stage, or specific message hierarchy. The result is generic creative that could serve to anyone at any stage of the funnel, which means it is not particularly well-suited to any of them.

Treating creative as a programmatic variable, not an afterthought, means briefing it with the same rigour you apply to audience strategy. What stage of the funnel is this serving? What does this person already know about the brand? What is the single thing we want them to think, feel, or do after seeing this? Those questions should be answered before anyone opens a design tool.

Programmatic and the Broader GTM Stack

Programmatic does not operate in isolation. It sits within a broader go-to-market system that includes paid search, social advertising, organic channels, email, sales, and whatever offline touchpoints are relevant to the business. The question of how programmatic connects to that system is one that most organisations answer poorly, if at all.

The most common failure mode is channel siloing. Programmatic is managed by one team or agency. Paid search is managed by another. Social is managed by a third. Each is optimised independently, each reports its own numbers, and nobody has a clear view of how the channels interact. The result is duplication, gaps, and attribution conflicts where multiple channels claim credit for the same conversion.

A connected approach means thinking about programmatic’s role in the customer experience relative to other channels. Programmatic is generally better at reach and awareness than search, which captures existing intent rather than creating it. It is better at targeting defined audience segments than most social platforms, though the gap has narrowed. It is better at contextual placement and premium inventory access than many alternatives. Understanding these relative strengths, rather than treating every channel as interchangeable, is what allows you to allocate budget in a way that reflects how customers actually move toward a purchase.

The go-to-market thinking that underpins effective programmatic use is the same thinking that underpins effective marketing more broadly. Who are you trying to reach? What do you want them to think or do? How does this channel contribute to that outcome relative to the alternatives? Those questions sound basic because they are. The problem is that programmatic’s complexity, the platforms, the data, the auction mechanics, the measurement tools, tends to absorb attention that should be spent on the strategic layer above it.

If you want to think about how programmatic connects to a full growth strategy across channels and customer lifecycle stages, the articles in the Go-To-Market and Growth Strategy hub cover the broader framework in more detail. The channel-level decisions only make sense once the strategic layer is clear.

There are also useful external perspectives worth reading. Vidyard’s piece on why GTM feels harder captures something real about the increasing complexity of reaching buyers across fragmented environments, which is exactly the problem programmatic is supposed to solve. And if you are thinking about the tooling and growth infrastructure that sits alongside programmatic, Semrush’s overview of growth tools provides useful context on where media buying sits within a broader stack.

The Practical Setup: What a Functional Programmatic Operation Looks Like

For marketers who are building or auditing a programmatic setup, the following is a reasonable baseline for what functional looks like, as distinct from what sophisticated looks like. Most organisations are not at sophisticated yet, and trying to skip the baseline is a reliable way to waste money on complexity you are not ready to use.

A functional programmatic operation has clear campaign objectives that are linked to business outcomes, not just platform metrics. It has a DSP relationship that provides transparency into where spend is going at the domain and app level. It has brand safety controls that are actively configured and periodically reviewed, not set once and forgotten. It has viewability thresholds that are enforced, not just reported. It has audience segments that are based on first-party data or well-reasoned proxies, not just platform defaults. And it has creative that is specific to the funnel stage and audience, not generic.

Beyond that baseline, the sophistication layers include incrementality testing, advanced audience modelling, cross-channel frequency management, connected TV integration, and dynamic creative at scale. These are worth pursuing, but they build on the baseline, not around it.

One practical point that often gets overlooked: the people managing your programmatic setup matter more than the platform they use. The major DSPs are more similar than their sales teams would like you to believe. The difference between a well-run DV360 account and a poorly run one is not the platform. It is the expertise and attention of the people operating it. Invest in that before you invest in platform upgrades.

When I was growing the agency team at iProspect, one of the things that separated the accounts that performed consistently from the ones that lurched between good and bad quarters was the quality of the strategic thinking sitting above the execution. The platform operators who understood the commercial context of what they were doing, who understood why the client cared about the metrics they cared about, consistently made better decisions than the ones who were just managing bids. Programmatic is no different. The technology is a tool. The thinking is the product.

For broader thinking on how go-to-market strategy connects to growth execution, Forrester’s intelligent growth model is worth reading as a framework for thinking about where media investment sits within a commercial growth architecture. And BCG’s work on scaling agile is relevant if you are thinking about how to build the internal capability to manage programmatic at scale, since the organisational design questions are as important as the technical ones. CrazyEgg’s overview of growth approaches also provides useful framing on how paid media channels, including programmatic, connect to broader growth mechanics.

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