SEO forecasting your finance director will believe

8 min read

SEO forecasting your finance director will believe

8 min read

Finance directors have seen plenty of SEO forecasts. Most show a line rising steadily to the right, built on total market search volumes and an assumption that the brand will rank first for everything. They are approved reluctantly, or not at all, and quietly forgotten when the line is missed.

SEO forecasting can be done in a way finance will accept. The method is not complicated: the demand you can realistically win, multiplied by the share of clicks you can expect, your own conversion rate and your own order value. What makes it credible is that every input is a stated assumption, the answer is a range, and the forecast is tracked against actuals every month.

This article sets out how we build an SEO forecast and business case, the mistakes that undermine one, and how to report it once the budget is approved.

Why the forecast matters more than the strategy

The forecast is the part of an SEO business case the board actually reads. It sets the budget, the expectation and the date by which someone will be asked whether it worked.

Overstate it and you win the budget once, then lose credibility when results land below the line. Understate it and the programme is funded too thinly to move anything. A forecast with stated assumptions and a range avoids both, because when the numbers move, everyone can see which assumption was wrong.

The model in one line

Forecast contribution equals search demand in scope, multiplied by achievable click share, multiplied by conversion rate, multiplied by average order value, multiplied by margin. For lead generation, replace order value with lead-to-customer rate and customer value.

Each of those is an assumption the business fills in from its own data. This is where each one comes from and who should own it:

Demand in scope: keyword research tools, checked against your own Search Console impressions. Owned by your SEO lead or agency.
Achievable click share: your own Search Console click-through rates by position, adjusted for what appears on the results page. Owned by your SEO lead or agency, and challenged by you.
Conversion rate: your analytics, for organic landing pages of the same type. Owned by e-commerce or analytics.
Average order value and margin: your order data. Owned by finance.
Costs: the agency quote plus internal time, content and development. Owned by marketing.

When finance supplies the commercial inputs, the forecast stops being the agency's number and becomes the business's. The outcome measure changes by model, which our playbooks for e-commerce, B2B lead generation and SaaS each set out.

SEO forecasting step by step: how we build one

These are the eight steps we follow for every forecast, in this order, because each one depends on the decisions made before it.

Define the scope as the non-brand terms you will target. We group terms by the page or template that will rank for them (categories, products, guides) and leave brand searches out, because people searching for you by name will mostly find you anyway. Search Console's branded queries filter makes the current split easier to see.
Size demand honestly. Search volumes from keyword tools are estimates, so we use 12 months to capture seasonality and cross-check against your Search Console impressions for terms you already appear for. The forecast covers only terms the site has pages for, or will build, never the whole market.
Set click share as a stated assumption, not an industry curve. Published click-through rate curves average across very different results pages. We use your own Search Console data for similar queries at each position (Search Console also counts clicks on links in AI Mode), then adjust for what sits above the organic results: ads, Shopping listings and AI answers.
Allow for AI answers taking clicks. Pew Research Center found that users clicked a traditional result in 8% of visits with an AI summary against 15% without one, in a study of 900 US adults' browsing in March 2025. That is one study in one market, so we treat it as a reason to apply a lower click assumption to queries that show AI Overviews, not as a figure to copy. It is part of the wider move to zero-click search.
Model the ramp over time. Rankings do not arrive on day one. We set out, quarter by quarter, when each group of terms is expected to move, tied to the work: technical fixes when developers release them, new pages when they are published, harder terms as authority builds. If release dates slip, the forecast slips with them, and the report says so.
Convert to revenue with your own data. We use your organic conversion rate by landing page type, because category pages and guides convert differently, with your own average order value and margin. Where your data allows, we split new and returning customers, because a board values a new customer differently.
Build three scenarios. Low, mid and high cases vary the inputs that are genuinely uncertain, usually click share and timing, and hold the rest at your own historical figures. Multiplying every optimistic assumption together produces a high case nobody should believe.
Add cost and payback. We plot cumulative cost against cumulative contribution in each scenario and show the month the lines cross. If the low case does not pay back within the period finance cares about, we say so, because that is a decision for the board and not something to bury.

What to put in front of the board

The forecast that goes to the board should fit on one or two pages, with the working model available to anyone in finance who wants to test it. We include:

A scenario table: low, mid and high cases for clicks, revenue and contribution by quarter, with no figure shown more precisely than the inputs allow.
The assumption list: every input, its value in each scenario, its source and its owner.
The payback chart: cumulative cost and cumulative contribution for each scenario, with the crossing point marked.
What is excluded: brand demand, terms without a planned page, and any market or channel left out of scope.
The risks: what would push results towards the low case, such as delayed releases, a core update or more AI answers on key queries.

Time to impact and what finance should expect

Organic search pays back later than paid media, because the cost comes first and the return builds. Finance should expect a period where cost runs ahead of contribution, and the forecast should show how long that lasts in each scenario rather than avoid the question.

Google's own guidance on hiring an SEO says that no one can guarantee a #1 ranking on Google. A forecast should not imply one either. Its job is to show the shape of the return, the range around it and the assumptions that drive it.

How to track the forecast against actuals

A forecast is only believed the second time if the first one was tracked. We report against it monthly at the level of each input, not just the total:

Impressions against forecast demand, to see whether the market moved.
Clicks and click-through rate against the click share assumption.
Conversion rate and order value against the commercial assumptions.
Release dates against the ramp: was the work shipped when the forecast assumed it would be?

When the total misses, this shows why: the market shrank, the results page changed, the site converted worse or the work shipped late. Each has a different fix. Google says sites appearing in AI Overviews and AI Mode are counted in the Search Console Performance report, so falling clicks with steady impressions points to the results page rather than the site. We re-forecast quarterly and keep the original alongside, so the board sees both.

Mistakes that make finance stop believing

Using whole-market volumes rather than the terms in scope.
Including brand terms, which inflates the forecast with revenue you already have.
Assuming first position for every term, or one generic click-through curve for every query.
Ignoring AI answers and other features that take clicks before the organic results.
One number with no range, given to the nearest pound, which implies a precision nobody has.
Ignoring seasonality, so a seasonal peak looks like progress and the trough looks like failure.
Assuming work the team has no capacity to ship.

When a forecast does not fit

Some decisions need a different document. A new product category with little search demand needs a case for creating demand, not a forecast of capturing it. A site migration needs a forecast of risk: what could be lost and how to protect it.

Where the question is whether to enter a market or channel at all, a market feasibility report comes first and the forecast follows from it. The cost side should match what the scope needs, which our guide to what drives SEO pricing explains.

Questions to ask your team or agency

Which terms are in scope, and are any of them brand searches?
Where does each click-through assumption come from, and how does it allow for AI answers?
Which inputs came from our own data, and which are estimates?
What would have to be true for the high case, and how likely is that?
How will we report forecast against actual, input by input?

How we forecast

We build forecasts this way for every organic search programme we propose, with the assumptions written down for finance to challenge. If you need a business case your board will approve, get in touch.

Frequently asked questions

How do you forecast SEO traffic and revenue?

Take the search demand for the non-brand terms you will target, multiply it by the share of clicks you can realistically win at each expected position, then apply your own conversion rate, average order value and margin. Use your own Search Console click-through data rather than a generic curve, phase the gains over time, and give low, mid and high scenarios with every assumption written down.

How accurate are SEO forecasts?

No SEO forecast is precise, because rankings, results page layouts and demand all change. A useful forecast gives a range rather than a single number, states every assumption, and is tracked against actual results each month at the level of each input. That way, when results differ from the forecast, you can see whether demand, click share, conversion or delivery was the cause.

Should an SEO forecast include branded search?

No. People searching for your brand by name mostly find you whether or not you invest in SEO, so including branded demand inflates the forecast with revenue you already have. Forecast non-brand terms only, and use Search Console's branded queries filter or a brand term filter to set the baseline. Report brand search separately as a measure of brand demand.

How do AI Overviews affect an SEO forecast?

AI answers can reduce clicks on queries where they appear. A Pew Research Center study of 900 US adults in March 2025 found clicks on traditional results in 8% of visits with an AI summary, against 15% without one. Treat this as a reason to use lower click assumptions for affected queries, and track AI visibility alongside clicks.

How long does SEO take to pay back?

It depends on the gap between your site and competitors, how quickly developers can ship changes and how competitive the terms are. A good forecast shows cumulative cost against cumulative contribution for low, mid and high scenarios and marks the payback month in each. Costs come first and returns build, so expect a period where cost runs ahead of return.