Advertising Trading: What the Buy Side Never Tells You

Advertising trading is the process by which media inventory is bought and sold between advertisers, agencies, and publishers, either directly or through automated systems like programmatic platforms. It determines not just where your ads appear, but at what price, under what conditions, and with what degree of transparency about what you are actually getting.

Most marketers engage with advertising trading every day without fully understanding the mechanics underneath it. That gap between surface familiarity and structural understanding is where budget quietly disappears.

I spent years on the agency side managing significant media budgets across multiple markets. One of the more uncomfortable things I learned is that the people spending your money often have financial incentives that do not perfectly align with your outcomes. That is not a conspiracy. It is just the structure of the industry. Understanding advertising trading, at a commercial level, is one of the few things that genuinely shifts the power dynamic back toward the client.

What Is Advertising Trading and How Does It Actually Work?

At its most basic, advertising trading is the exchange of media inventory for money. A publisher has space, an advertiser wants to reach an audience, and some mechanism brings them together. That mechanism has evolved substantially over the past two decades, and the complexity introduced along the way has not always served advertisers well.

The three main trading models are direct deals, programmatic guaranteed, and open auction programmatic. Each sits at a different point on the spectrum between control and efficiency.

Direct deals involve a negotiated agreement between an advertiser or agency and a publisher. You agree a price, a volume, a placement, and a set of conditions. The upside is transparency and control. The downside is that it requires relationships, lead time, and minimum commitments that do not suit every campaign.

Programmatic guaranteed sits in the middle. You negotiate terms directly but fulfil the buy through automated systems. You get some of the control of a direct deal with some of the operational efficiency of programmatic. It is increasingly common for premium inventory that publishers do not want to push into the open market.

Open auction programmatic, often called real-time bidding, is where most digital display, video, and connected TV inventory is traded today. Advertisers bid for impressions in milliseconds through a demand-side platform, or DSP. Publishers make inventory available through a supply-side platform, or SSP. Ad exchanges sit in between, matching bids to supply. The whole chain sounds elegant. In practice, it is layered with intermediaries, each taking a margin, and the cumulative effect on working media is significant.

If you want to understand the commercial mechanics of how trading fits into broader go-to-market planning, the Go-To-Market and Growth Strategy hub covers the strategic context in which these decisions sit.

Where Does the Money Actually Go?

This is the question most clients never ask precisely enough, and most agencies never answer precisely enough.

In a programmatic supply chain, the advertiser pays a gross CPM. By the time that impression is served, a DSP has taken a fee, an ad exchange has taken a cut, an SSP has taken a margin, and the publisher receives what is left. Independent research into programmatic supply chains has consistently found that publishers receive a minority of what advertisers spend. The exact percentage varies, but the direction is always the same: less reaches the publisher than most clients assume.

This matters for two reasons. First, if your working media is lower than you think, your effective CPM is higher than your reported CPM. Your cost per impression, per click, per conversion, all of it is being calculated on a number that flatters the channel. Second, lower publisher revenue tends to correlate with lower quality inventory. Premium publishers who can sustain direct deals or programmatic guaranteed arrangements do not need to flood the open market. What ends up in open auction is often inventory that could not be sold any other way.

I have sat in client meetings where the agency presented a blended CPM that looked competitive, and the client was satisfied. What was not discussed was how much of that spend was clearing through the agency’s own trading desk, at a margin the client had not explicitly approved. This is not illegal. It is often disclosed somewhere in a contract the client signed without reading carefully. But it is a structural conflict of interest, and it is worth knowing about.

What Are Trading Desks and Why Do They Create Tension?

Agency trading desks emerged in the early programmatic era as a way for holding groups to centralise media buying and capture more of the value chain. The proposition to clients was efficiency and scale. The reality was more complicated.

A trading desk sits between the client’s budget and the open market. The agency buys inventory at wholesale rates, marks it up, and charges the client a retail price. The spread is the agency’s margin. In some arrangements, the client pays a transparent fee on top of the actual media cost. In others, the margin is embedded in the CPM and the client never sees the underlying cost.

The tension arises because the trading desk has an incentive to route spend through inventory where its margin is highest, which is not always the inventory where your campaign performs best. When I was running an agency, I saw this tension play out regularly. The performance team wanted to optimise for client outcomes. The commercial team had volume commitments to inventory partners. Those two objectives do not always point in the same direction.

Independent trading desks and in-house programmatic operations have grown partly as a response to this problem. Brands that have brought programmatic buying in-house report a clearer picture of where money goes, though the operational complexity is real and the talent required is not cheap.

BCG’s work on commercial transformation in go-to-market strategy makes a relevant point: companies that understand their cost structures and where value leaks are consistently better positioned to make strategic investments. The same logic applies to media trading. You cannot optimise what you cannot see.

How Does Auction Dynamics Affect What You Pay?

Real-time bidding operates on auction logic, and understanding that logic changes how you think about bidding strategy, budget pacing, and campaign setup.

Most programmatic auctions today run on a second-price model, meaning the winning bidder pays one cent more than the second-highest bid, not their actual maximum bid. This was the dominant model for years. However, many exchanges have shifted toward first-price auctions, where the winner pays their actual bid. The shift happened gradually and was not always communicated clearly to advertisers. If your bidding strategy was calibrated for second-price dynamics and the auction moved to first-price, you may have been overpaying without knowing it.

Bid shading emerged as a partial solution. DSPs began using algorithms to estimate the clearing price and submit bids slightly below the maximum, attempting to reduce overpayment in first-price environments. The problem is that bid shading is a black box. You are trusting the DSP’s algorithm to act in your interest, and the DSP has its own commercial relationships with exchanges that may influence how aggressively it shades.

Floor prices add another layer. Publishers set minimum CPMs below which they will not sell inventory. Floor prices are often dynamic and are sometimes set in ways that push auction clearing prices artificially upward. This practice, sometimes called price manipulation, has attracted regulatory attention in several markets. The point for advertisers is that the price you pay in a programmatic auction is shaped by forces you often cannot observe directly.

Earlier in my career, I was very focused on driving down CPMs as a measure of trading efficiency. I thought lower cost meant better performance. What I eventually understood is that CPM is a unit cost, not a value measure. A cheap impression that reaches nobody useful is worse than an expensive impression that reaches the right person at the right moment. The obsession with CPM optimisation is a symptom of measuring what is easy to measure rather than what matters.

What Is the Difference Between Efficiency and Effectiveness in Trading?

This is where advertising trading connects to the broader strategic debate about performance marketing, and it is a distinction worth being precise about.

Efficiency in trading means getting more impressions, clicks, or conversions for less money. Effectiveness means those impressions, clicks, or conversions are actually moving your business forward. The two can diverge significantly.

Programmatic optimisation is very good at efficiency. Algorithms will find the cheapest path to a conversion event. The problem is that conversion events are a proxy for business outcomes, not the outcomes themselves. If your conversion event is a form fill, the algorithm will find the cheapest form fills. Whether those form fills become customers, and whether those customers are profitable, is a different question that the algorithm does not answer.

I have seen this play out in practice more times than I can count. A campaign hits its cost-per-acquisition target. Everyone is pleased. Then someone looks at the downstream data and finds that the leads are converting at a fraction of the rate of leads from other channels. The algorithm was efficient. It was not effective. The difference was invisible until someone looked past the trading metrics.

The same logic applies to audience targeting. Programmatic platforms are very good at finding people who look like your existing customers. That is retargeting and lookalike modelling. What it does not do particularly well is reach genuinely new audiences who have no prior signal of intent. For growth, you need both. Capturing existing demand is not the same as creating new demand, and an advertising trading strategy that optimises only for the former will eventually run out of runway. The mechanics of market penetration require reaching beyond your current base, and trading strategy needs to reflect that.

How Should You Evaluate Inventory Quality?

Not all impressions are equal, and the programmatic ecosystem has a persistent problem with inventory quality that advertisers need to take seriously.

Invalid traffic, or IVT, refers to impressions generated by bots or other non-human sources. Estimates of the proportion of programmatic traffic that is invalid vary, but the consensus is that it is material. You are almost certainly paying for some non-human traffic in your programmatic campaigns. The question is how much, and what you are doing about it.

Brand safety is a related issue. Open auction programmatic can place your ads next to content that conflicts with your brand values or creates reputational risk. Inclusion lists, exclusion lists, and contextual targeting tools exist to manage this, but they require active management. Default settings on most DSPs are not optimised for brand safety. They are optimised for scale.

Viewability is another dimension. An impression is only valuable if a human has a reasonable chance of seeing it. An ad that loads below the fold on a page a user never scrolls to is an impression in the reporting but not in any meaningful sense. Viewability standards exist, but compliance is inconsistent across inventory sources.

When I was growing an agency from a small team to over a hundred people, one of the things we invested in early was ad verification tooling. Not because clients were demanding it, but because we knew that without it, we could not make honest claims about what we were delivering. The clients who asked the right questions about inventory quality got better outcomes. The ones who only looked at headline metrics were often getting less than they thought.

What Does a Smarter Trading Approach Look Like in Practice?

Understanding the mechanics of advertising trading is not an academic exercise. It has direct implications for how you brief agencies, structure contracts, evaluate performance, and allocate budget.

Start with transparency. Before you can evaluate whether your trading is working, you need to know what you are actually buying. That means asking for log-level data, not just aggregated reports. It means understanding what fees are embedded in your media costs versus charged separately. It means knowing whether your spend is going through a trading desk and what the commercial arrangement is.

BCG’s research on pricing strategy in go-to-market contexts is instructive here: understanding cost structures at a granular level is a prerequisite for making good commercial decisions. The same principle applies to media. You cannot negotiate well from ignorance.

Second, align your trading strategy with your growth objective. If you are trying to defend market share among existing customers, retargeting and high-intent search make sense. If you are trying to grow your customer base, you need trading strategies that reach people outside your current data set. That typically means investing in higher-funnel inventory, which is harder to attribute but not less valuable. Vidyard’s research on pipeline and revenue potential for go-to-market teams points to a consistent finding: untapped pipeline requires reaching people who are not yet in your system.

Third, think carefully about your measurement framework before you set your trading strategy. The metrics you optimise for in-platform will shape what the algorithm does. If you optimise for click-through rate, you will get clicks. If you optimise for view-through conversions with a long attribution window, you will get credit for conversions that would have happened anyway. Neither is the same as optimising for profitable customer acquisition. Set your success metrics at the business level first, then work backward to what in-platform signals are honest proxies for those outcomes.

Fourth, consider the channel mix question seriously. Programmatic display is one trading environment. Paid search is another. Social platforms operate their own walled garden auctions. Connected TV is a different market again. Each has different cost dynamics, audience characteristics, and attribution behaviours. A trading strategy that treats all channels as interchangeable will make poor allocation decisions. Tools that help you understand growth across channels can support this, but the strategic logic has to come first.

Finally, do not assume that more automation means less oversight is needed. The opposite is true. Programmatic systems move fast and optimise toward whatever signal you give them. If that signal is wrong, or if the data feeding it is poor, the system will efficiently do the wrong thing at scale. Human judgment, applied at the strategy and setup level, is what keeps automated trading aligned with actual business objectives.

The broader strategic context for these decisions sits within go-to-market planning. If you are working through how advertising trading connects to your overall growth model, the Go-To-Market and Growth Strategy hub is worth spending time with. Trading mechanics are only useful when they serve a coherent commercial strategy.

What Should Marketers Ask Their Agencies Right Now?

If you manage an agency relationship and advertising trading is part of the scope, there are a handful of questions that will tell you a lot about how well the arrangement is actually working for you.

Ask what percentage of your media spend reaches the publisher. If the answer is vague or the agency cannot produce a clear number, that is informative. Ask whether any of your spend passes through the agency’s own trading desk or affiliated entities, and if so, what the commercial terms are. Ask to see the inclusion list for your programmatic campaigns and the logic behind it. Ask how invalid traffic is measured and what the threshold for concern is.

These are not hostile questions. They are the questions a commercially literate client should be asking. Agencies that are doing good work and operating transparently will have clear answers. The ones that deflect or respond with jargon are telling you something.

When I was on the agency side, the clients I respected most were the ones who understood enough about the mechanics to ask sharp questions. They made us better. They forced rigour. The clients who just looked at the dashboard and said the numbers looked good were the ones we were, honestly, less worried about. Not because we were doing anything wrong, but because scrutiny sharpens performance. It always does.

Understanding how advertising trading works is not about distrust. It is about being a sophisticated buyer in a complex market. The more you know about how inventory is priced, how auctions work, and where the money flows, the better your trading decisions will be and the more useful your agency conversations will become. Growth strategies that incorporate creator-led approaches, like those explored in Later’s work on go-to-market with creators, also require a clear-eyed view of how media inventory is traded to ensure those investments are structured effectively.

Advertising trading is not glamorous. It does not generate conference talks or award entries. But it is where a significant portion of marketing budget lives or dies, and treating it as a back-office function rather than a strategic one is a mistake that compounds quietly over time.

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