How to Do SEO Forecasting for SaaS Without Making Up Numbers

SEO forecasting estimates what a planned set of SEO work could bring in, usually as a range of clicks, demos and pipeline over a period. A useful forecast for a SaaS company is built from your own conversion data and shows its assumptions, so anyone can challenge them. It can't predict exactly what Google will do, and a forecast presenting one precise number is claiming more than anyone knows.

You'd do this to make a decision, such as whether to fund a content program or hire someone. So the test of a forecast is whether it helps with that decision, and not whether it turns out to be correct to the dollar.

Start by choosing the question

Different decisions need different forecasts.

• "Should we invest in SEO at all?" A rough range, showing whether the opportunity is large enough to matter.
• "Which of these two projects should we do first?" A comparison, where only the relative size is important.
• "How much pipeline should we plan on next year?" A cautious baseline, plus a separate estimate for planned work.

Say you run a 30-person SaaS for managing clinical trials, with a sales-led motion. Leadership asks whether building 30 commercial pages is worth the cost of a content lead. You don't need a model with a hundred inputs. You need a sensible low, middle and high case.

Step 1: Build a keyword-based opportunity estimate

For new pages, work from demand.

1. List target phrases, grouped by planned page. Use your keyword map.
2. Get volumes from a keyword tool. Tools are estimates, and they're particularly unreliable for small B2B phrases. So treat these as rough, and note that real demand can be different in either direction.
3. Decide what position you might realistically reach. Look at who ranks now. How to read keyword difficulty on a new site can help you with this. Don't assume the top spot for everything.
4. Apply a click-through rate for that position. Don't borrow a generic industry curve, if you can avoid it. Your own Search Console figures reveal the click rates you get at different positions, on your own kinds of page. Calculate them from pages similar to the ones you're planning.

Then you can multiply: monthly volume, times the expected click rate at the position you think you can reach. Add it up across pages.

Results pages have changed, so adjust for that. Search the phrases, and note any AI Overviews, ads and other features taking up the space above the organic results. Informational phrases with an AI Overview may deserve a much lower click rate than the same position once earned. Commercial phrases may hold up better. You can test this with your own data as well. See a zero-click search strategy for SaaS.

Step 2: Use your own history if you have it

If you've published comparable pages before, your own history is better evidence than any tool.

Look at your last 10 or 20 similar pages. How long did each take to get its first meaningful clicks? What did they level off at? What share never performed at all? That last number matters, and it's usually left out. If a third of your past articles got almost nothing, a forecast assuming that every page succeeds isn't realistic.

If you have an existing trend, you can also project it forward as a baseline, which shows what happens with no new work. Allow for seasonality by comparing with the same months last year. Then estimate the new work separately on top. Keeping the two apart tells you what the investment itself is expected to add.

Step 3: Convert clicks into business numbers

Traffic alone won't help anybody decide. Use conversion rates from your own analytics and CRM.

Step	Example calculation (illustrative)
Monthly clicks to target pages	2,000
Visit-to-demo rate on similar pages	Your own number, by page type
Demo-to-opportunity rate	From your CRM
Opportunity win rate	From your CRM
Average first-year contract value	From your CRM

Split it by page type where you can, because a comparison page and an educational blog post convert very differently. If you don't have enough data for any one step, say so, and use a wider range. Our guide to measuring SEO ROI for SaaS goes further into these calculations.

Don't forget to include timing, either. Assume that pages take several months to ramp up, as new ones rarely perform on day one. Then add your sales cycle on top. For the clinical trials SaaS, a page published in January might start bringing in demos by spring and closed deals in the second half of the year, or later. A forecast that ignores this can make a good project look like it's failed.

Step 4: Present a range, with the assumptions showing

Create three cases, changing only a few inputs:

Input	Low	Middle	High
Share of pages reaching target position	Few	Some	Most
Average position reached	Lower	Mid	Higher
Click rate	Reduced for AI Overviews	Your own average	Your better pages
Ramp-up time	Slow	Typical	Fast

Put the assumptions on the same page as the result. It's far better for a CFO to say "I think your win rate is too high" than to accept a number that nobody can inspect.

Compare the result to the cost. Include the salaries, freelancers, tools and any engineering time. Then ask the useful question: how much of the low case would we need for this to pay off? If even the low case is fine, go ahead. If only the high case works, think again, or consider a smaller test.

What can't be forecast

Be open about what isn't in the model:

• Google updates and SERP changes, which can move things in either direction
• Competitors, who may publish, improve or outspend you
• AI answers, which may reduce clicks for some queries, and also influence buyers who never click
• Brand effects, such as a buyer who reads your post and books a demo through a direct visit months later

So the number you're forecasting is the part you can see. Some real value won't show up, and some forecast clicks won't arrive. Say both.

There's no reliable way to forecast AI citations at the moment, and you should be wary of anyone who offers one. It's better to report it separately, using a monthly prompt test, without trying to turn it into a dollar figure.

Forecasts that should worry you

If someone hands you an SEO forecast, whether from an agency or a tool, check the following:

• One exact number, with no range.
• The assumption of top positions for most target phrases.
• Generic click curves, with no allowance for what the results page looks like now.
• Traffic as the only outcome, with nothing on conversions.
• No ramp-up time, where the results begin in month one.
• A guarantee. Nobody controls Google, and Google itself warns that no one can guarantee a top ranking.

Compare the forecast against reality

A forecast should have a review date. After a few months, compare what happened against what you assumed, one input at a time. Did the pages get indexed on schedule? Did they reach the expected positions? Were click rates in line? Did conversions match?

This lets you find out which assumption was wrong, and your next forecast will be more accurate as a result. Going back to the clinical trials SaaS, it might learn that its comparison pages ranked more slowly than expected but converted far better. That's more useful than knowing that the total was a little off. It also directly changes what to build next. Track it using the KPIs that belong in your report.

If you'd like a realistic estimate for your own category, book a call with Comppound.