Quality Path Optimization: Going Beyond Cheapest-Path SPO

In programmatic, the same ad slot on the same page can reach your DSP through several routes. The publisher might sell it through three SSPs, and some of those SSPs may receive it from resellers. Each route takes a cut, and each has its own risk of misrepresentation. SPO was the industry's answer to that duplication. Quality path optimization is a refinement of it.

What quality path optimization means

The term is not a formal standard. Vendors and agencies use it in slightly different ways. The common idea is that a path should be judged on two questions:

1. Is it efficient? Does more of the dollar reach the publisher, with fewer hops?
2. Is it good? Are the impressions authorized, human, viewable, and on inventory you would choose?

Cheapest-path SPO answers the first question well. It can answer the second one badly. A path can be cheap because it carries remnant inventory, mislabeled domains, or traffic that would fail an IVT check. If you only optimize on price, you may route more spend into exactly those paths.

How it differs from classic SPO

The existing supply path optimization guide for pharma covers the classic version: what to remove and why. This page is about the quality layer on top.

The signals quality path optimization uses

ads.txt

Publishers list the sellers authorized to sell their inventory in an ads.txt file (app-ads.txt for apps). If a path claims to sell a domain but the seller ID is not in that file, the path is unauthorized. Also note whether the seller is listed as DIRECT or RESELLER.

sellers.json

SSPs and exchanges publish a sellers.json file identifying the entities they pay. It lets you see whether a seller ID belongs to the publisher, an intermediary, or is marked confidential. A path that ends in a confidential or unidentified seller deserves a closer look.

SupplyChain object (schain)

The schain object in the bid request lists every node the impression passed through. More nodes usually means more fee take and more chance of misrepresentation. Many buyers prefer one or two hops and treat long chains as a warning. The article on using ads.txt, sellers.json, and schain together explains how to read all three.

IVT and viewability by path

Verification data split by path, rather than by campaign alone, shows which routes carry bot traffic or low viewability. A path with IVT several times the campaign average is a candidate for removal even if it is cheap. The explainer on viewability, attention, and IVT covers how these are measured.

MFA and attention signals

Some buyers add flags for made-for-advertising sites and attention scores. These are less standardized. Use them to rank paths, not as pass or fail gates, unless you trust the method.

How to implement it

1. Get path-level data. You need impressions, spend, and verification results by SSP, seller ID, and domain. Log-level data helps a lot here; see what log-level data to request.
2. Check authorization. Flag any spend where the seller is not authorized in ads.txt or cannot be identified in sellers.json.
3. Measure hops. Group spend by schain length. Look at quality metrics by group.
4. Score each path. Combine fee take, IVT, viewability, and MFA share into a simple score. Keep the method simple enough that someone can explain it in a QBR.
5. Cut in stages. Remove the worst paths first. Re-check reach against your HCP target list or DTC audience after each round.
6. Protect unique supply. If a path is the only route to an endemic publisher your HCPs use, keep it and fix the issue with the seller instead.
7. Re-run quarterly. Paths change. New resellers appear. A clean path last spring may not be clean now.

What usually goes wrong

The most common failure is cutting too hard on a small audience. A rare-disease HCP list may only reach a few thousand people, and removing paths can drop reach without anyone noticing until the reach report comes back. Measure reach on the list, not impressions.

The second failure is trusting a vendor score you cannot inspect. If a "quality score" moves spend between paths, you should be able to see what went into it.

The third is doing this once. A quality path list built in Q1 and never revisited will drift. The programmatic media quality guide suggests folding this into a quarterly review.

Practical takeaway

Ask for last month's spend grouped by schain length (one hop, two hops, three or more), with IVT and viewability for each group. If the longer chains are carrying a real share of spend and posting worse numbers, you have your first quality path cut.

Editorial note. Analysis and frameworks are the author's own and do not represent Acxiom or any current or former employer, client, or named platform. Examples labeled hypothetical or illustrative are not results from real campaigns. Nothing here is legal, regulatory, or medical advice.