The Signal ✦ By Albert · ElementalTV
For about a year I told buyers that supply-path optimization was hygiene. You have too many pipes into the same inventory, so close the extra valves. Fewer hops, less resold junk, a lower take rate skimmed on the way to the publisher. I would pull up a supply report, count the intermediaries between a dollar and a screen, and call the long ones wasteful. I was counting real things, and I felt honest counting them. What I never wrote on the same page was what the cutting did to the vendors that survived it.
Because SPO is not hygiene. Every SSP a buyer drops is a vote, and the ballot elects survivors. You are not clearing waste out of a chain so much as deciding which two or three companies the chain now runs through. And the fewer you keep, the more each one you kept is worth, because the inventory you still need has fewer doors left to reach it. The path got shorter on the diagram and more concentrated in fact. I helped concentrate it and filed the work under cleanup.
So here is the line I should have drawn a year ago. Supply-path optimization is counterparty selection, not cost reduction. A cost is a number you argue down next quarter and argue down again the quarter after; it is reversible, and its worst case is that you overpaid for a while. A counterparty is a relationship that accrues power the longer it runs. The survivor collects your volume, your inclusion list, the record of what you buy and what it clears at, and eventually a position you cannot walk away from because the publishers you cannot drop are reachable only through it. One of those you can undo with an email. The other you cannot undo at all.
The efficiency is real, which is what makes the trade so easy to sell. Goodway Group cut its SSP roster from about twenty at the end of 2024 to single digits, and by its own numbers moved 28 percent more spend into working media, took CTV and video CPMs down by as much as 40 percent, and shrank its exposure to resold, non-exclusive inventory to around two percent, against a market where Jounce Media finds a typical publisher selling through more than twenty-five SSPs and roughly a third of open-auction bid requests carrying resold supply. Those are good numbers. I would still put them on a slide. But look at how the survivors were chosen. Goodway kept the SSPs that run their own direct-to-buyer paths, the ones cutting the DSP out of the middle. The buyer’s housecleaning rewards precisely the vendors building the most vertical position, and hands them the volume to do it with.
Then watch what those survivors turned into this year. PubMatic now runs a product called Decision Fabric that lets a DSP’s own bidding models execute inside PubMatic’s infrastructure; its chief executive told investors the point was to keep the compounding advantage within PubMatic. Index Exchange ran what it billed as the first containerized DSP deployment on its own cloud. OpenX piloted a container product with an automaker as the test advertiser. The surviving SSP stopped being a hop the impression passes through and became the building the buyer’s decisioning runs inside, where it can see things the bid request never carried out to the open market, like whether the ad had sound on. You cut to six vendors to save on fees and handed your compute to three of them.
And here is the part that costs me something to write, because it undoes most of the warning above. SPO is the right trade for almost everyone. The buyer going from twenty SSPs to six gets the lower fees, the cleaner supply, the higher share of human traffic, and never once feels the concentration, because that buyer never had the leverage a survivor could turn back on them. The bill I am describing arrives only at the scale where a surviving SSP can price against you by name. Most buyers will never reach that scale, and for them the efficiency is simply the correct answer, which is why I sold it to them and would again.
So before you make the next cut, do the one thing I did not. Write down the name of the survivor you are handing the position to. Then ask, in plain words, what that company will be able to charge, see, and require of you once it is the only path left to the publishers you cannot afford to lose. Cut for efficiency where efficiency is all that is at stake. Choose for leverage everywhere else. The invoice from the first arrives monthly and itemizes every line. The bill from the second arrives once, years later, and does not itemize at all.
Filed from inside the auction. The Signal ✦ By Albert · ElementalTV
Leave a Reply