Flat Monthly Rates vs CPM: The Case for Selling Time

A flat monthly rate sells a named position for a period of time. A CPM deal sells a promised number of impressions. If your site is small, selling the month is safer for both sides, because a quiet month costs nothing to explain.

Most publishers reach for CPM first because it is what the industry talks about and it feels like the grown up option. It is also a promise, and the size of a promise you should make depends entirely on how predictable the thing you are promising is.

What is the actual difference?

A flat monthly deal says: this slot on these pages is yours for December, for $250. Whatever traffic arrives is what your ad gets. A CPM deal says: you are buying 40,000 impressions at $6 per thousand, for $240. If 40,000 do not arrive, you have not been given what you paid for.

That second sentence is the whole difference. A CPM deal creates a delivery obligation, and when it is not met the remedy is a make good: extra time, extra impressions or money back. On a large site that is routine housekeeping. On a site with one sidebar and forty thousand views, it is a personal email to somebody you know explaining that your traffic went down.

Why is a guaranteed count risky when you are small?

Because small site traffic is lumpy in ways nothing you do controls. One post reaches the front page of a forum and the month doubles. A search update lands and the month drops a fifth. A big site averages that out across hundreds of advertisers and thousands of pages. You have one slot, one advertiser, and one month.

Here is the same slot, three months running, sold both ways.

Month	Flat rate at $250	Slot views	CPM at $6
March	$250	41,000	$246
April	$250	52,000	$312
May	$250	33,000	$198
Quarter	$750	126,000	$756

Six dollars apart over three months. The difference that matters is not in the money column: it is that the CPM version needed a conversation in May and the flat version did not.

May is the month worth looking at twice. Under the flat deal, nothing happens. The advertiser paid $250, their ad ran in the position they bought all month, and a slightly quieter month is simply what May was. Under the CPM deal you have delivered 33,000 of a promised 40,000, and now you owe a week of free delivery in June, or a credit note, or an explanation. You have lost $52 of revenue and gained an admin task, and the advertiser has had a slightly worse experience of dealing with you.

You are not selling impressions. You are selling a position on a page that people read.

Which should you sell, and when?

Notice that most of the right hand column describes a site with an ad operation, and most of the left hand column describes a person with a website. Sell the model that matches which of those you are this year, not the one that matches what you hope to be.

Do you still tell them how many impressions to expect?

Yes, always, and you use the word estimate. Something like: this slot has averaged about 40,000 views a month over the last three months, though it is an estimate rather than a guarantee, and I will send you the actual figures at the end of the month.

That sentence does three useful things at once. It gives the buyer the number they need to work out whether the price is sane. It sets the expectation that the number moves. And it promises a report, which is the single most effective thing you can do to make a first month turn into a second one. Publishers who go quiet after the invoice are the ones who find out about a non renewal by silence, which is the pattern in why advertisers stop replying.

Report actual delivery every month either way. Selling time instead of impressions is a decision about what you promise, not a licence to stop counting.

How do you work out the flat number?

Start from the CPM you would have charged, multiply by a realistic month, and round to something a human can remember. A slot averaging 40,000 views at a $6 CPM gives you $240, so the flat price is $250. Rounding up slightly is fair: the advertiser is getting the certainty of a fixed cost and the guarantee that nobody else is in that position.

Two guard rails. Do not set the flat price below what the same inventory would earn you as network remnant, because then you are paying for the privilege of doing sales work. And if the slot also runs in an auction when it is unsold, keep your floor price consistent with the flat rate, or you will undercut yourself automatically while you sleep. The pricing groundwork for both of those is in how to price ad space on your website.

Who should not sell flat?

Three cases, honestly.

• Strongly seasonal sites. If December is four times November, one flat rate across both is wrong twice. Price the seasons separately or sell those months on CPM.
• Anyone selling to agencies. A media buyer working across thirty sites needs comparable units, and flat monthly deals do not slot into their spreadsheet. Give them impressions.
• Slots you intend to share. If several advertisers are bidding for the same position and splitting delivery between them, you are already running an impression based system and a flat exclusive month contradicts it. Pick one per slot: either it is sold whole for the month, or it is in the auction.

For everybody else, and that is most people running one site, the flat month wins on the things that actually decide whether direct selling works: it is easy to quote, easy to invoice, easy to renew, and it never once requires you to apologise for the weather. Once the site is big enough that the arithmetic starts costing you real money, the whole comparison across every pricing model is in CPM, CPC or CPA: which should you sell.