Display advertising is being rebuilt from the ground up. Instead of a salesperson negotiating a placement on a website for a month, software now buys individual impressions one at a time, deciding in the milliseconds while a page loads whether this particular person is worth showing an ad to. That is programmatic advertising. It is genuinely a better mechanism, and it has introduced a set of problems that the industry is only starting to admit exist.
What actually happens when a page loads
The sequence is worth understanding, because every problem in this article follows from it. A visitor opens a page. The publisher’s ad server sends a request describing the slot and whatever is known about the visitor. Advertisers, or rather the demand-side platforms bidding on their behalf, evaluate that request against their targeting and submit a bid. The auction resolves and the winning ad is delivered. All of it completes in roughly 100 milliseconds.
The appeal is obvious. You are no longer buying "a banner on this site for a week," you are buying "this person, right now, who has visited a product page in the last fortnight." Budget flows toward impressions that are worth something instead of being spread evenly across a placement.
The questions nobody asks the vendor
Because the process is automated and invisible, it is unusually easy to spend money without knowing what you bought. Three issues deserve attention before you commit budget.
Was the ad ever actually visible?
An impression is recorded when an ad is served, not when a human sees it. An ad loaded at the bottom of a page nobody scrolled to still counts. This gap is large enough that the industry built a standard specifically to address it: through the Making Measurement Make Sense initiative launched in 2011 by the IAB, ANA, and 4As, the Media Rating Council defined a viewable display impression as at least 50% of the ad’s pixels in the viewport for at least one continuous second, and for video, 50% of pixels for at least two continuous seconds.
50% / 1s
MRC standard for a viewable display impression
50% / 2s
MRC standard for a viewable video impression
~100ms
Time for a real-time auction to complete
Source: Media Rating Council viewability standards, developed through the 3MS initiative.
Read that standard carefully, because it is a floor rather than an achievement. Half an ad, visible for one second, counts as viewable. That is a long way from "someone read it." Buy on viewable impressions rather than served impressions, and understand that even viewable is a generous definition.
Was it a person?
Automated buying created an automated incentive to fake the supply. Fraudulent operations generate non-human traffic to sites built solely to carry advertising, and the money is real even when the audience is not. Ask what invalid traffic detection your provider uses and what proportion of spend it filters.
Where did the money go?
Between your budget and the publisher sit a demand-side platform, an exchange, a supply-side platform, a data provider, and often a verification vendor. Each takes a percentage. The share of your spend that reaches the publisher who showed the ad is considerably smaller than most advertisers assume, and it is frequently not disclosed. Ask for it in writing.
The auction optimizes for the cheapest impression matching your criteria. Cheap impressions are cheap for a reason, and that reason is rarely in your favor.
What this means for a small business
Programmatic rewards scale and data, which is precisely what a small advertiser lacks. A campaign with a modest budget spread across the open exchange tends to buy exactly the impressions nobody else wanted. That is not an argument against using it, but it is an argument for using it narrowly.
| Use case | Verdict | Why |
|---|---|---|
| Retargeting site visitors | Strong | You supply the audience data, so you are not relying on the exchange to find relevance. Highest return available here. |
| Reaching a customer list | Strong | Same logic. Your own data is better than anything you can buy. |
| Broad prospecting on the open exchange | Weak at small budgets | You are competing against advertisers with vastly more data, and you will win the impressions they declined. |
| Brand awareness with no targeting data | Poor | Pay for viewable impressions on named premium inventory instead, where you know what you bought. |
Key takeaways
- Programmatic buys individual impressions at auction in roughly 100 milliseconds, rather than fixed placements.
- A served impression is not a seen one. The MRC standard is 50% of pixels for 1 second (display) or 2 seconds (video), and that is a floor.
- Automated buying created an incentive to fake supply, so ask what invalid traffic filtering is applied.
- Multiple intermediaries take a cut. Ask in writing what share of spend reaches the publisher.
- For small advertisers the reliable uses are retargeting and first-party audiences, not open-exchange prospecting.
Related reading
Sources
- Programmatic
- Strategy
Written by
Valter Brandt
Chief Marketing Officer
Valter Brandt is the Chief Marketing Officer of ThisCom, working with clients across the United States and internationally. He has led marketing strategy through the major shifts in social advertising, mobile, content marketing, programmatic media, and marketing automation.
All articles by Valter Brandt →
