Auction mechanics underneath every exchange used to buy and sell adult traffic

Last updated: 7 September 2026

Real-time bidding on 18+ inventory follows the same protocol as mainstream programmatic and then diverges in most of the details that actually matter. A page loads, the publisher sends a bid request, demand partners respond inside a fixed window, and the winning creative renders on the page. Every decision available to either party happens inside that window. Understanding what travels in the request and what determines settlement separates a deliberate effort to buy and sell adult traffic from the habit of funding an account and hoping for the best.

What a bid request carries before either party can buy and sell adult traffic

A bid request is a structured message describing an impression that has not happened yet, and its completeness decides what anyone can achieve when they buy and sell adult traffic programmatically. Fields left empty by a publisher never get filled in later by anybody, so an incomplete request permanently limits what the demand side can decide. Completeness is the constraint.

It carries the placement identifier, format and dimensions, a device and connection description, a coarse geographic signal and a floor price. Everything a demand partner knows at decision time arrives inside that message. Domain disclosure is where adult supply differs most from mainstream, because many publishers here withhold it and demand sees an identifier where a site name would otherwise sit.

Inferring quality without a domain name

Behavioural signals substitute for identity. Consistency of delivery across hours, the ratio of sessions to reported clicks, and stability of conversion rate over several weeks separate real placements from filler more reliably than any single request field. None of them requires cooperation from the publisher to collect.

Build the whitelist on those signals and then keep it. Identifiers stay stable inside an account, so four weeks of testing produces an asset that continues working long after the testing budget has been spent. That asset is the closest thing to a durable advantage available on this inventory, and it cannot be bought from anybody at any price.

Floors, settlement and the price you actually pay to buy and sell adult traffic

A floor is the minimum a publisher will accept and it is set per placement rather than per account. Reading floors before bidding saves a week for anyone new to how exchanges let participants buy and sell adult traffic. Floors are published in most panels and ignored by most newcomers, which is why so many first campaigns report no delivery whatsoever.

Bids below the floor never enter the auction at all, which is why a campaign can report almost no delivery while appearing correctly configured everywhere else in the panel. Settlement rules vary and they matter. Under second-price settlement the winner pays slightly above the runner-up rather than its own bid, so a higher bid buys position without automatically buying a higher price.

First-price settlement removes that cushion completely and changes strategy with it. There the bid is the price, overbidding is pure loss, and calibration against observed clearing prices becomes the entire discipline rather than a refinement of it. Confirm which model applies before setting a ceiling, since panels rarely state it clearly.

Why floors move without an announcement

Publishers adjust floors in response to their own fill and to seasonal demand, usually through automated tools rather than by hand, and nothing notifies the demand side when it happens. Automated floor management reacts to fill within hours rather than weeks, so movement can arrive overnight and reverse just as quickly. Watch the share rather than the volume.

A placement that cleared comfortably last month can sit above your ceiling this month with nothing having changed on your side of the transaction. Watch impression share against bid rather than delivered volume alone. A sudden fall in share with a stable bid indicates a floor movement, while a fall in both usually indicates an inventory problem, and the two need opposite responses.

Mediation, backfill and the second demand layer for publishers who buy and sell adult traffic

Unsold impressions pay nothing, and mediation exists to reduce how many of them there are. Any publisher intending to buy and sell adult traffic profitably runs at least two demand sources for that reason alone. A single demand source leaves inventory unsold at exactly the hours when the audience is largest.

A mediation layer calls demand sources in sequence until one bids above the floor, and it is standard practice on any established 18+ property. Sequencing matters more than the number of partners involved, because calling a high-paying but low-fill source first and a reliable filler second captures the premium without losing the impression to an empty slot on the page.

The latency cost of a long waterfall

Each additional call in a sequence adds delay before anything renders, and delay costs impressions on mobile connections where users leave first. Three or four sources is usually the practical limit. Each call in the chain has to time out before the next begins, and the accumulated delay is paid on every impression rather than only on the ones that fail.

Measure render rate rather than fill rate when tuning a waterfall, because fill counts what sold while render counts what a person could actually have seen, and only the second number turns into revenue by the close of a billing month. The difference between them widens with every extra partner added to the chain, which is why longer waterfalls stop paying for themselves.

Reporting on both sides and the discrepancies that follow when you buy and sell adult traffic

Buy-side and sell-side reports never match, and the mismatch is structural rather than dishonest. Anyone who watches both while continuing to buy and sell adult traffic learns to read the gap as information. The gap is stable enough to predict once measured, which turns a recurring argument with a supply partner into an agreed adjustment.

Impressions counted at bid time exceed impressions counted at render, which exceed impressions a person could have seen. Each system counts at a different point in the sequence and each is internally consistent. My reconciliation habit came from a reporting reference for desks that buy porn traffic while monetising their own placements, and it ended a long argument with a supply partner.

Which number to use for which decision

Use rendered impressions for revenue, bid-level impressions for auction diagnostics, and recorded sessions for anything touching conversion rate. Mixing them produces ratios that look alarming and mean nothing at all. I borrowed the convention from an analytics team that has to buy adult traffic on mixed inventory. Pick one definition per metric, write it into the report header, and refuse to compare across definitions afterwards.

Keep the definition written down beside the report. Six months later nobody remembers which counting point a historical figure used, and a comparison across two definitions is worse than no comparison, since it produces a confident answer pointing in the wrong direction entirely.

Pacing, bid strategy and the controls open to desks that buy and sell adult traffic

Accelerated pacing spends a daily budget as fast as auctions allow, concentrating delivery into the cheapest hours rather than the best ones. Switching it is the first control worth using, a point taken from a pacing note written for traders who buy and sell adult traffic on floor-heavy inventory. The default costs nothing to change and changes everything downstream of it.

Even pacing spreads delivery across the day and produces a sample worth reading afterwards. Per-source bidding is the second control and it is not universal across platforms. Without it every placement inside a campaign shares one bid, so converting sources are underbid and dead ones overbid at the same moment, on the same line, in the same report.

Calibrating a bid against observed clearing prices

Set the opening bid slightly above the published floor and raise it in small increments while watching impression share. The point where share stops improving is the ceiling worth keeping, and bidding past it purchases nothing except a higher average price. Increments also make the ceiling visible rather than theoretical.

My increment discipline began as a scribbled rule from a trader working thin margins. Publishers tune floors and waterfalls while buyers tune bids and whitelists. Participants who buy and sell adult traffic well treat the auction as a mechanism to be measured rather than a box to be funded. Both sides are reading the same auction.