What it means
Advertisers want to place messages in suitable contexts, and publishers want to sell available space. Programmatic systems connect rules, prices and delivery across many placements.
The IAB describes programmatic terminology and OpenRTB, and Google documents programmatic guaranteed deals, which shows that real-time bidding is one method, not the whole category. A fictional retailer sets a campaign to reach local shoppers on selected sites, and software checks available impressions against its budget and rules while a human still chooses the objective and reviews results.
Demand-side platforms help buyers manage campaigns, supply-side systems help publishers offer inventory, and exchanges or marketplaces can connect them, with variations in how deals are agreed. A fictional publisher reserves premium homepage space for one advertiser under a guaranteed agreement; delivery uses software, but the inventory is not won by an open auction, so calling it ordinary RTB would be inaccurate.
In real-time bidding, an impression opportunity can be offered to eligible buyers and a bid selected quickly, with the platform and deal rules determining eligibility and price. Not every bid wins or becomes a viewable ad.
A fictional brand bids for a mobile placement, loses the auction and never displays its ad, because a bid request is not an impression delivered to a person. Targeting may use contextual signals, first-party data or other permitted information, but privacy rules and platform changes restrict certain uses, and more data is not automatically better or lawful.
A fictional campaign targets content about running shoes rather than building detailed individual profiles, and the team checks audience fit and consent requirements instead of assuming a data vendor's promise settles compliance. Brand safety and suitability controls help avoid unwanted contexts but are imperfect, as a fictional education brand found when its ads appeared beside misleading content and it had to adjust exclusions, question the supplier and report where ads actually appeared.
Invalid traffic, ad fraud and poor viewability can waste spend, and independent verification or platform controls may help because a served impression is not always human attention. A fictional advertiser buys one million reported impressions, verification flags some as invalid or not viewable, and the team calculates cost against valid delivery rather than using the headline alone.
Fees can sit at several points in the supply chain, so a fictional buyer who spends 10,000 through a platform separates media from service fees before comparing campaigns, because a low bid price may not mean low total cost. Frequency controls try to limit how often the same person sees an ad but may not work perfectly across devices or platforms, so a fictional commuter seeing the same ad repeatedly on several apps is a cue to review frequency and reach rather than raise bids.
Creative and landing pages still matter, since automation can choose placements but cannot fix a misleading offer such as a fictional travel ad promising a fare no longer available. Measurement should match the objective and recognise attribution limits: a fictional retailer with high attributed sales found through a holdout that many shoppers would have bought anyway, so it changed targeting to improve incremental results, because programmatic advertising automates transactions, not judgment.
In practice
Real-world examples.
Example
A demand-side platform bids on eligible display impressions.
Example
A publisher sells a programmatic guaranteed homepage placement.
Example
A brand reviews invalid traffic and actual ad placements.
Formula
Calculation
Effective CPM = total relevant ad spend / valid delivered impressions x 1,000, using a stated definition of valid delivery.
Worked example. A fictional advertiser spends $12,000 on a campaign that reports 1,000,000 impressions, of which verification finds 800,000 valid and viewable.
- Reported CPM = $12,000 / 1,000,000 x 1,000 = $12.
- Effective CPM = $12,000 / 800,000 x 1,000 = $15.
The $3 difference shows the cost of paying for delivery that did not count, which is why the stated definition of valid delivery matters.Case study
Seen in the real world.
In this fictional case, Alder Shoes runs a programmatic campaign across selected publishers. It excludes unsuitable contexts and sets a frequency goal. After launch, verification finds poor viewability on some placements. The team shifts spend and assesses incremental sales rather than reporting bids alone.
The invented finance lead also separates media cost from platform and data fees before comparing the campaign with other channels. That comparison shows which channel produced sales at an acceptable cost and informs the next quarter's budget. The figures and company are invented for illustration.
Watch out
Common mistakes.
- Treating all programmatic buys as open real-time auctions.
- Equating bid requests or served ads with human attention.
- Ignoring fees, privacy and placement quality.
Questions
People also ask.
Is programmatic advertising the same as RTB?
No. RTB is one programmatic buying method.
Does software remove the need for people?
No. People set strategy, creative, controls and review.
What should an advertiser check?
Inventory, valid delivery, fees, data permissions and outcomes.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%Related