SEO forecasting done honestly: how to model traffic and revenue from rankings you do not have yet, which inputs matter, and how to hold a forecast accountable.
Every SEO proposal should answer one question before it asks for budget: if this works, what is it worth? SEO forecasting exists to answer that, and done honestly it is one of the most useful exercises in marketing. Done dishonestly, it is a hockey-stick chart drawn backwards from the number someone wanted to hear.
What a forecast actually models
An SEO forecast is a chain of assumptions multiplied together: the keywords you could realistically rank for, the search volume they carry, the click-through rate at the positions you might achieve, your conversion rate from that traffic, and your value per conversion. None of those inputs is exotic. The honesty lives in how each one is estimated.
Keyword-level volume and difficulty come from tools and from the SERPs themselves. Click-through rates by position are published and reasonably stable, though features like AI overviews are reshaping them, which is why we model conservative and optimistic curves rather than one number. Conversion rate and order value should come from your own analytics, never from industry averages.
The three mistakes that make forecasts fiction
First, assuming rankings arrive instantly. Real forecasts are time-phased: little in the first quarter, momentum in the second and third, compounding after that. Anyone forecasting month-two returns from a standing start is selling something. Second, forecasting the head terms only. Most SEO revenue arrives through long-tail queries that individually look too small to model, so good forecasts include a modelled tail rather than pretending the five big keywords are the whole market. Third, ignoring the competitive response: your rivals are not frozen while you climb.
Ranges, not points
We present forecasts as scenarios: a conservative case built on defensible assumptions, an expected case, and an upside case that says what happens if the content earns links faster than planned. A single-point forecast is a guess wearing a suit. A range is a decision-making tool, because it lets a finance director see the downside case and still say yes, or no, with clear eyes. It is the same logic we apply in incrementality testing: marketing numbers earn trust by admitting their uncertainty.
Holding the forecast accountable
A forecast is only useful if you revisit it. Ours become quarterly scorecards: which assumptions held, which did not, and what that changes about the plan. Rankings arriving faster than modelled but converting worse is a different problem from rankings arriving slowly, and the forecast is what lets you tell them apart. This is the CLEAR Method applied to organic search: commercial measurement built in from the start, not reconstructed at renewal time.
Forecasting also disciplines the investment conversation itself. When you know what a realistic programme returns and when, questions like whether a fee is reasonable stop being about the fee and start being about the return on it.
If you are weighing an SEO investment now
Ask whoever is proposing it, us included, to show the assumptions, not just the chart. Which keywords, what positions, what click-through curve, whose conversion data. If the assumptions survive scrutiny, the forecast is worth acting on. If they cannot be produced, you have learned something more valuable than any forecast.
Talk to us if you want a forecast built on your actual market and your actual numbers. Our SEO team would rather lose a pitch on honest numbers than win one on fiction, because only one of those clients is still with us in year four.
