Direct Deals vs Open Marketplace — When to Go Private
So for the first four or five months of running pop traffic as a publisher I had every single placement set to public. Meaning literally any buyer on PopLayer could see my inventory, link a campaign to it, and start bidding. I thought this was the obvious move — more buyers seeing my traffic means more competition means higher CPMs, right? Econ 101. Supply and demand. I felt very smart about this.
Except thats not what was happening. What was actually happening, and I only figured this out because I got bored one Tuesday evening and started really digging into my per-placement stats, was that my best placement — 15,000 daily impressions, mostly US desktop, genuinely good traffic from a content site I'd been building for two years — was being won by the same buyer every single day. Same guy, same campaign, paying my floor CPM. Every. Single. Auction. Because he was the only one bidding. I had essentially set up a private deal with one buyer at the worst possible price without meaning to. He was getting US desktop traffic for $1.80 CPM and probably laughing all the way to whatever his conversion rate was.
That Tuesday evening I switched the placement to private. Within a week I had reached out to three buyers I knew were running offers in the US market (you can kind of figure out who wants what by looking at the marketplace activity for a while). All three linked campaigns. Now I had four buyers competing for the same inventory including the original guy. Average CPM went from $1.80 to about $3.20 within ten days. Same traffic, same placement, same everything. Just more people bidding on it.
Forty percent revenue increase because I changed one dropdown from "public" to "private" and sent three messages. I'm still a little mad I waited five months to do it.
When staying public actually makes more sense though
Before you go and switch everything to private let me save you from a mistake I also made (yes there were many mistakes, its a whole theme with me). I got excited after the US desktop win and switched ALL my placements to private. Including my little sites that do like 2,000-3,000 impressions a day of mixed global traffic. Nobody requested access to those. Why would they? There are a million sources of mixed-geo traffic at 3k daily volume, theres nothing special about mine specifically. Those placements earned exactly $0 for about two weeks until I switched them back to public.
Public placements are the right call when youre new and nobody knows your site exists yet — you need to be discoverable. Also good for commodity traffic (mixed geos, moderate volume, nothing remarkable about it). The auction will find a reasonable price and you dont have to do anything. I keep four of my smaller sites on public right now and they bring in steady revenue with zero maintenance from me. Its not exciting money but its consistent and I dont have to think about it, which is worth something.
Theres a middle option too — "request only" where the placement shows up in the marketplace but buyers have to request access before they can bid. I use this for placements where the traffic is decent but not premium enough to go fully private. Buyers can find me but I still get to vet them before they start running campaigns. Kind of the best of both worlds if you dont mind checking your access requests every few days.
The negotiation thing that I was afraid of but turned out to be easy
When a buyer requests access to your private placement you have leverage and this is the part where I see most publishers mess up because they either dont negotiate at all (just approve and leave the floor at whatever default they set) or they set the floor so high that nobody bids and then wonder why their earnings are zero.
What I do is check what the buyer is running. Direct response campaigns (CPA offers, app installs, lead gen) can generally afford higher CPMs than brand awareness stuff because the buyer is optimizing for conversions and a converting user is worth real money to them. Then I look at my own RPM stats for the geos this buyer seems to care about. If my US traffic runs at $4 RPM theres absolutely no reason to set a floor at $1.50. I start at 70-80% of what I think the traffic is worth and leave a little room so the buyer feels like theyre getting a fair deal. Then I watch the fill rate for a week — if its below 50% the floor might be too high, if its above 80% I probably left money on the table and could push it higher.
What I actually run today and what it earns
Hybrid setup. Four smaller sites with mixed traffic stay public — they average about $1.20 CPM across all geos combined. Two sites with strong US/UK traffic have private placements with 2-4 approved buyers each — those average about $2.90 CPM. The private placements represent maybe 30% of my total impression volume but they generate close to half my total revenue. Maintaining them takes maybe 20 minutes a week to check access requests and review buyer campaigns, which is a pretty good hourly rate when you think about it.
Quick list of things I learned the hard way so you can skip that part: dont go private until you have at least 5,000 daily impressions because below that nobody will bother requesting access. Always check what URL the buyer is actually sending traffic to before you approve them — I approved one buyer without looking and they ran some sketchy thing that got one of my domains flagged by Google Safe Browsing, which was a fun week of cleanup. And even in private deals, set different floor CPMs per geo. A buyer will happily pay $3 for US traffic but they're not paying $3 for India, and if your single floor is $2 you lose the US buyer entirely while also being too expensive for the Indian traffic. Per-geo floors, always.