Most publishers fail at this by trying to sell before they can measure, or by measuring forever and never selling. Here is an order that avoids both.
Weeks 1–2: measure
Install an ad server and create one placement in your strongest position — below the header or mid-content. Fill it with a house ad for your own newsletter.
Do nothing else. You are collecting a baseline: impressions per day, and the click rate a genuinely relevant ad achieves in that position.
What you should have at the end: two weeks of impression data and a CTR figure.
Weeks 3–4: build the two assets
A rate card. One page. Who reads the site in one specific sentence, real impression numbers, the price, the sizes you accept, and how to book.
A prospect list. Twenty names, from three places: brands already advertising in your subject elsewhere, brands your readers mention, and any affiliate partners who already convert for you.
What you should have: a live /advertise page and twenty names with contact emails.
Weeks 5–8: send twenty emails
Five a week, personalised. Reference something specific about their product and why this audience fits.
Expect: 20 sent, 4–6 replies, 1–2 who want to talk, 1 who buys. That is a normal outcome, not a bad one.
When you get a reply, offer the two-week test. It converts far better than a monthly commitment and gives you a case study either way.
What you should have: one paying sponsor, or five conversations in progress.
Weeks 9–12: deliver, report, renew
The renewal is worth more than the sale, and it is decided by what you do now.
Send a report in week two of the flight, unprompted. Impressions, clicks, CTR, and any conversions.
Send a proper report at the end, with a specific ask:
19,400 impressions, 121 clicks, 0.62% CTR, and 6 orders attributed to the ad. Want to keep the slot for next month?
If it underperformed, say so and offer something. A sponsor told the truth and offered a better position will usually try again. A sponsor who has to work out for themselves that it did not work is gone.
What you should have at day 90: one renewed sponsor, a real case study, and a defensible price.
The numbers that decide everything
Three, and you only need three:
- Monthly impressions on the slot — what you are selling
- CTR — whether the position works
- Conversions attributed to the ad — the only number that renews sponsors
The third is the one most publishers skip, and the one that turns a $300 flight into a $300-a-month relationship. If the sponsor can see revenue, price stops being the conversation.
What not to do in the first ninety days
Do not add more slots. One slot that sells beats four that sit empty, and empty slots make the site look unsold.
Do not build a sales deck. A one-page rate card is enough at this scale, and nobody at a company small enough to buy from you is reviewing a deck.
Do not chase big brands. A national advertiser will not buy 20,000 impressions. Aim at companies where one person can decide to spend $300.
Do not lower the price to close the first deal at any cost. Shorten the commitment instead. The number you accept first is the number you renew at.
After ninety days
If you have one renewing sponsor, repeat the whole thing for a second slot. If you have none after twenty properly personalised emails, the problem is usually one of three things: the audience is too general for anyone to see themselves in it, the price is far off, or the prospect list was not really qualified.
All three are fixable, and all three are worth knowing after ninety days rather than after a year.