Floor Prices: The Number That Stops You Being Undercut

A floor price is the lowest amount you will accept for a thousand views of one ad slot. Any bid below it is refused outright.

It is the most consequential number you set, and the mistake almost everybody makes on their first day is setting it too high.

It is what you refuse, not what you get

This is the part worth understanding properly, because it changes how the number feels.

A floor does not fix your price. It removes the bottom of the market. When two advertisers both want a slot they bid against each other above the floor, and you get the higher number. So the floor is where your earnings start rather than where they stop.

Where to get the number

From what you already earn. Look at what your current ad network pays you per thousand views on that slot, then price above it.

That single rule makes direct sales arithmetically safe. If your floor is above your network rate, then any impression sold directly is worth more than the impression it displaced, and the ones you do not sell fall through to the network anyway. See how the passback works.

Your situation	Try	Why
Network pays you under $1	$1.50 to $3.00	Even a modest direct rate is several times a network rate.
Network pays you $1 to $3	$4.00 to $8.00	You have a niche worth something. Price like it.
Narrow professional audience	$10.00 to $25.00	The advertiser has nowhere else to reach these people.
You have no network and no data	$2.00	Start there, and let a month of results tell you the rest.

These are starting points for a first month, not valuations. The only real answer comes from whether the slot sells.

Too high earns nothing, not less

This is the failure worth spelling out, because it does not feel like a failure. A slot priced above what anybody will pay does not sell slightly less often. It does not sell.

Same traffic, same audience. Only the floor changed.

The cheap floor wins because a slot that sells every time earns around the clock while an expensive one sits empty.

An empty slot at four dollars pays exactly as well as no slot at all, and it does it while looking like you are holding a firm line.

Reading the answer a month later

Three things you might see, and what each one means.

• Almost no impressions sold. The floor is above the market. Halve it and wait another week. This is the most common outcome and the easiest to fix.
• Plenty sold, and your eCPM is sitting exactly on your floor. One advertiser is interested and nobody is bidding them up. You can try raising it, and be ready to put it back.
• Plenty sold, and your eCPM is comfortably above your floor. Several buyers are competing. Leave it alone. This is the state you want and the one a nervous publisher breaks by getting greedy.

Different slots, different floors

Floors are per slot, and that is deliberate rather than fiddly. A leaderboard at the top of your busiest page and a box at the bottom of an archive page are not the same product and should not be the same price.

A reasonable first spread: your highest floor on anything above the content, the middle on sidebars and in-article slots, and your lowest on anything below the fold or on a quiet page. Something is better than nothing down there, and pricing a weak slot like a strong one is how you end up with a page full of empty boxes.

When to change it

Give any change a full week before judging it, and change one slot at a time. Daily numbers bounce around enough that a single day tells you nothing, and two changes in one week tell you nothing about either.