SEO Forecasting: What the Models Get Wrong

Every SEO forecast is a chain of estimates multiplied together. Understanding which link is weakest tells you how much to trust the number.

What a forecast is

Almost every SEO forecast, whether from a platform or a spreadsheet, is the same calculation. Take a keyword search volume. Multiply by an expected click-through rate for a target position. Multiply by an assumed conversion rate. Multiply by average order value. Sum across keywords.

Four estimates multiplied together. If each is off by 30 percent in the same direction, the final number is off by roughly three times. That is why forecasts presented to two decimal places should make you suspicious.

Forecasting still has value. Precision is just the wrong thing to ask of it.

The search volume problem

Search volume figures come from keyword tools, which derive them from clickstream data and platform APIs, then round and bucket them. They are monthly averages, often over twelve months, which means a term with heavy seasonality reads as flat.

For local businesses the volumes are frequently too small for the tools to report accurately at all. A term showing 10 to 100 searches a month might be 4 or might be 90. Building a revenue forecast on that range is guesswork with extra steps.

Your own Search Console impression data is more reliable than any third-party volume estimate for terms you already appear for, because it is measured rather than modelled.

The click-through rate curve is not universal

Forecasts apply a standard CTR curve: position one gets roughly 30 percent, position two around 15, and so on. Those curves come from aggregate studies across all query types.

Your queries are not average. A query with an AI overview, a featured snippet, a map pack, and four ads above the first organic result has a completely different curve. Position one on that page might earn 8 percent, a long way from the 30 the curve assumes.

Branded queries run far above the curve. Definitional queries run far below it. A forecast that applies one curve to all of them will overstate informational terms and understate branded ones, which is precisely backwards for a business trying to decide where to invest.

Nobody models the competition responding

The deepest flaw. Forecasts assume the search results hold still while you improve. They do not. Your competitors are also publishing, also building links, and also responding when they lose position.

Google also changes the layout. A category that had ten blue links two years ago now has an AI overview and a shopping carousel above them. The organic click opportunity shrank regardless of anything you did.

The practical implication is that forecasts get less reliable the further out they go, and twelve-month projections in competitive categories are closer to narrative than prediction.

How to build one you can defend

Use ranges. "Somewhere between 400 and 900 additional sessions a month by month nine" is honest and still actionable. A single number implies a confidence nobody has.

Anchor on your own data. Use your measured click-through rate at your current positions in place of a generic curve, and your measured conversion rate in place of an industry benchmark.

Forecast the near term only. Three to six months is defensible. Beyond that, state assumptions instead of numbers.

Model the downside. What happens if you rank at position eight rather than three. If the case only works at position one, it is not a plan.

The metric worth forecasting instead

Traffic forecasts get requested. Revenue forecasts get funded. But the number that predicts both, and that you can measure early, is qualified impression growth in the query set you care about.

If impressions are rising on commercially relevant queries, position will follow, and clicks follow position. That signal appears within weeks, and it is measured, so you are not waiting quarters on an estimate.

It also fails fast, which is the useful property. If eight weeks of work produces no movement in relevant impressions, something is wrong with the plan, and you learn that before spending two more quarters.

What to ask a provider offering a forecast

Ask which click-through rate curve they used and whether it accounts for the SERP features on your actual queries, whether the volume figures are third-party estimates or your own Search Console data, what the forecast looks like if you land at position eight, and what happens to the number if a competitor responds.

A provider who has thought about it will answer directly and probably widen the range. One who has not will restate the original number more confidently. That response tells you what the forecast was for.