The campaign runs for weeks. AfterBid follows what happens after the auction: what was delivered, what it cost, how it performed, what changed, and what happened afterwards.
The auction is a moment. The campaign is a system.
AfterBid keeps the brief, the supply, the delivery and the outcome connected, so a later decision can still be reconstructed from evidence rather than memory.
The technology separates hard facts from interpretation, then keeps both attached to the same campaign record.
Publisher intelligence maps the page, the ad stack and the available supply signals: Prebid, GAM slots, ads.txt, identity, CMP, wrappers, bidders and the path through the market.
Execution telemetry adds what cleared, where it ran, what it cost, pacing, auction economics and placement-level delivery. Measurement joins when it becomes available.
Platform-specific connectors translate each source at the edge. Campaigns, deals, placements, creatives, measurements and decisions use one common model above them.
Deterministic rules protect hard gates. Specialist agents interpret the brief and evidence, explain the recommendation, then assign the action to the execution platform, the operator or the buyer.
Every intervention stores the evidence, the reason, the owner and the outcome. The next planning cycle begins with what was already learned instead of another blank dashboard.
Start with one execution platform. Add the next without rebuilding the intelligence above it. A new DSP, SSP, ad server or measurement source should add evidence, not force a new product. The campaign model, agents and decision history remain intact while the connector surface expands.
Each agent has a narrow remit. Together they turn the same campaign record into planning, qualification, diagnosis, recommendation and reporting.
Objective, audience, geography, format, constraints and KPI.
Environment, inventory, audience evidence and eligible supply.
MFA, invalid traffic, path integrity and other required checks.
Scale, floors, competition and auction constraints.
What is under-delivering, blocked or diverging from plan.
Judges placements against the KPI chosen in the campaign brief.
Increase, keep, reduce or suppress, with the reason attached.
Reconstructs what changed and whether the intervention worked.
AfterBid does not stop at “something is wrong”. It separates the part the media operator can change from the part the buyer or DSP controls, then keeps the evidence attached to both.
First implementation: AfterBid is being built into Maridion's curation workflow. The interface below uses representative campaign data.
The delivery gap has two owners: Maridion can narrow the effective clearing requirement on its side; the buyer controls the bid it sends.
A report can tell you what happened. The record below tells you what the system believed, what changed, who owned the action and what happened afterwards.
Page, supply and suitability gates passed.
Delivery evidence diverged from the qualified environment.
Suppress placement. Retain publisher. Reallocate eligible spend.
Operator accepted the change; execution rail updated.
Post-change KPI attached to the original decision.
Publisher crawling and campaign execution do different jobs. AfterBid keeps them separate so one component does not inherit access it does not need.
The common model should not inherit a vendor's schema as product truth.
A recommendation can show which source supports it and when that source was last refreshed.
If a rail exposes fewer signals, the system carries that gap forward instead of manufacturing certainty.
AfterBid can propose the change without pretending it controls a buyer-side setting that belongs elsewhere.
For agencies, curators and media businesses that want their operating intelligence to survive the platform they happen to use today.