How to pick a daily rate for an ad slot when you have no benchmark, what to change it to, and why per-day pricing does not need a traffic estimate.
Nobody tells you what to charge. Networks never had to answer this question because an auction answered it for them, so the moment you sell space yourself you face a blank field and no benchmark. This guide is a way to arrive at a number you can defend, and then a way to correct it.
Start from what a week is worth to a buyer, not from your traffic. Name a business that would want your readers, guess what they would happily pay to be visible to them for a week, and divide by seven. That figure is your daily rate. Most small sites land between $2 and $10 a day per slot.
This feels unrigorous and it is more reliable than the alternative. The traffic-based method — estimate impressions, apply an RPM you found in a forum — imports the assumptions of a market you are not in. A local business is not buying a thousand impressions; they are buying a week of being visible to the people who read the town's cycling blog, and they will judge the price against what a leaflet drop or a local paper ad costs, not against a CPM.
Position matters more than area, but both count. A leaderboard across the top of a desktop page is the most valuable slot most sites have; a small rectangle low in a sidebar is the least. As a rough internal ratio, before any other consideration:
| Slot | Relative rate | Why |
|---|---|---|
| Leaderboard, 728×90 | Highest on desktop | First thing seen, above the content |
| Billboard, 970×250 | Several times a leaderboard | Dominates the page; a takeover, not a banner |
| Half page, 300×600 | Above a rectangle | Occupies the whole visible sidebar |
| Medium rectangle, 300×250 | Your reference point | Fits anywhere; price the others against it |
| Mobile banner, 320×50 | Below a leaderboard | Small, but seen by most readers on most sites |
| Small rectangle, 200×150 | Lowest | Same dwell, less room to say anything |
If most of your readers are on a phone, invert the desktop instinct: a mobile banner that everyone sees is worth more than a leaderboard that only a third of your audience ever loads.
Two signals, and they take about a month each to read. If the slot has been visible and for sale for several weeks with no enquiry at all, the price is high for the market or the slot is not visible enough. If it sold within a day or two and has been continuously occupied since, it is low.
A third signal is more useful than either: what happens when a second advertiser wants the same slot. If people are queueing for it, the price is definitively too low, and raising it costs you nothing because demand already exceeds supply.
No — change it for new purchases and leave running advertisers where they are. The trust you are building is worth more than the difference, and an advertiser who feels the rate moved under them will not come back. A rate rise applies to whoever buys next, which is the normal way a price change works everywhere else.
The exception is a slot you have deliberately underpriced to get started. Say so at the time — "introductory rate for the first month" — and then the increase is expected rather than imposed.
Discount rather than give away, and put a number on it. A free slot has no price to anchor against, so the conversation about renewing starts from zero. A first week at half rate establishes both that the slot is worth paying for and what it is worth, which is what you actually need from the first sale.
Not for the billing, no. A day costs what you said a day costs, whether that day brought forty readers or four hundred. Traffic still matters — it is what makes an advertiser think the price is fair, and it is why they renew — but it is not what the invoice is made of, which is the structural difference between selling space and selling attention.
Be ready to talk about it honestly anyway. An advertiser who asks how many people read the site and gets an evasive answer will assume the worst. A small, specific, truthful number is far more persuasive than a vague large one.
Related: the earnings calculator, ad sizes at true scale, and how pricing works in the product.
Keep reading