What is programmatic advertising?
Programmatic can put your brand in front of the exact accounts you sell to, across websites, apps and streaming TV. It can also burn budget on impressions that would have converted anyway. The difference is setup and measurement.
How programmatic advertising works
Before programmatic, buying display ads meant calling publishers, negotiating rates and signing insertion orders. Programmatic replaces that with software and auctions.
Here’s the basic flow. A visitor opens a page on an industry news site. The site’s SSP sends the ad slot, along with details about the page and the visitor, to an ad exchange. DSPs representing many advertisers receive that request and decide whether to bid and how much. The highest bid wins, and the winning ad loads with the page. All of it happens while the page renders. This auction is called real-time bidding (RTB).
The parts of the programmatic stack
Demand-side platform
The advertiser’s buying tool. You set budgets, audiences, bids and creative, and the DSP bids across many exchanges at once.
Supply-side platform
The publisher’s selling tool. It offers ad slots to exchanges and DSPs and sets price floors.
Real-time bidding
The open auction where each impression sells to the highest bidder as the page loads.
Private marketplaces
Invite-only auctions where selected advertisers buy premium inventory from specific publishers, usually at a set floor price.
Connected TV
Ads on streaming TV apps and smart TVs, bought programmatically with digital targeting and reporting.
Data and targeting
First-party lists, contextual signals, account lists and lookalike audiences that tell the DSP who to bid on.
Targeting options
Programmatic gives you several ways to choose who sees your ads:
- First-party audiences: your CRM contacts, site visitors and customers, matched in the DSP. This is the most reliable data you have.
- Account-based targeting: a list of target companies, reached at the office or by IP and firmographic data. Useful for B2B with long sales cycles.
- Contextual targeting: ads placed next to content on a topic, such as engineering trade sites. It works without third-party cookies.
- Retargeting: ads to people who visited your site and left without converting.
- Lookalikes: new people who resemble your best customers.
Set frequency caps so the same person sees your ad a few times, and keep refreshing creative before it fatigues.
Measuring incrementality
Platform dashboards report view-through conversions: people who saw an ad and later bought. Many of those people were already going to buy. A retargeting campaign aimed at last week’s cart abandoners will look brilliant in its own report, whatever it actually added.
Incrementality testing answers the real question: how many sales happened because of the ads? The usual method is a holdout. You split your audience or your regions, show ads to one group and hold back the other, then compare conversion rates over a full sales cycle.
Formula: Incremental lift = (Test group conversion rate − Holdout conversion rate) ÷ Holdout conversion rate
Say your test regions convert at 2.4% and your holdout regions at 2.0%. The lift is 20%. Apply that to revenue in the test regions and you get the revenue the ads truly drove. Feed CRM outcomes back into the platform so you’re measuring pipeline and closed revenue, beyond clicks. Our guide to the attribution funnel covers how this fits with search and AI channels.
From our work
Here’s how we’d approach programmatic for a high-AOV ecommerce brand already running Google and Meta ads. First, we’d connect the store and CRM so purchase data flows back to every ad platform, using our CRM and tech integrations. Next, we’d start small with first-party retargeting and a connected TV test in a few regions, with matched holdout regions that get no programmatic spend.
We’d judge the test on blended revenue and new customers in the test regions vs the holdout, measured over a full purchase cycle. If the lift holds up, budget scales. If the lift is flat, the money goes back to search and social. Our paid media service runs this kind of test.
Related terms
Programmatic connects to most of paid media. See demand-side platform, supply-side platform, private marketplace, connected TV, incrementality, cost per mille, retargeting and first-party data. For search ads, see our PPC management page.
Frequently asked questions
What is an example of programmatic advertising?
A machinery maker uploads its list of target accounts to a DSP. When an engineer at one of those companies reads an industry news site, the DSP wins the auction for a banner on that page, and the engineer sees the machinery maker’s ad.
What is the difference between programmatic and display advertising?
Display is an ad format: banners and rich media on websites and apps. Programmatic is a way of buying ads. Most display ads are now bought programmatically, and programmatic also covers video, audio and connected TV.
Is Google Ads programmatic?
Partly. Google’s Display Network and Display & Video 360 buy ads through automated auctions, which is programmatic. Search ads also run on an automated auction, though most people use “programmatic” to mean display, video and CTV buying.
Is programmatic advertising good for B2B?
It can be, when you target by account list or first-party data and measure against pipeline. Broad audiences and click-based reporting tend to waste B2B budget, since the buying committee is small and the sales cycle is long.
Pay for the lift, skip the noise
We run paid media tied to CRM revenue data, so every channel, programmatic included, gets judged on the pipeline it adds.
