TL;DR: SEO forecasting is the process of predicting future organic search traffic and revenue with simple arithmetic: estimate search volume, click-through rate, conversion rate, and average deal value, then model the outcome. The formula is straightforward; the inputs are the hard part. Search volume is an estimate, click-through rate now swings by more than double depending on whether Google shows an AI Overview, and the timeline is the input people fight about. Forecast a range instead of a number, write the assumptions down where the client can see them, set a date to revise it, and tie the forecast to measurable benchmarks and the tools or templates you use to build it.

Nobody asks for an SEO forecast because they are curious. They ask because they are about to spend money and they want to know what comes back. That is true whether you are an SEO, a marketer, an agency client, a finance team trying to evaluate the investment, or a founder who needs a number that will survive a planning meeting.

I have built these for prospects, for clients arguing with their own finance team, and once for a founder who wanted a number for a board deck. That last one taught me the most. The number went into the deck and sat there for eleven months, unchanged, while reality moved around it. That is why good SEO forecasting is less about sounding precise than about setting realistic expectations you can revisit as rankings, SERP features, and conversion rates change.

Google’s documentation on hiring an SEO draws a line I keep coming back to. It says an SEO audit should be about giving you realistic estimates of improvement and an estimate of the work involved, and that if someone guarantees their changes will give you first place in search results, find someone else. Estimates are expected. Guarantees are a red flag. The gap between those two words is where this entire discipline lives, and it is the gap this guide addresses with the forecasting formula, the weak points in the inputs, and a practical way to build forecast ranges you can defend.

The formula, and what it hides

Keyword-level SEO forecasting is not complicated arithmetic. Take search volume, multiply by the click-through rate you expect at your target position, and you have projected traffic. Multiply that by conversion rate and average deal value and you have projected revenue.

Patrick Stox laid out that same formula in Ahrefs’ guide to the subject, with a caveat worth repeating. It only accounts for the keywords you typed in, and it tends to assume one keyword per page when pages actually rank for dozens. He makes another point I agree with more every year, which is that complex forecasts take considerably more effort and are not reliably better than simple ones.

So the math is easy. Every input feeding it is an estimate, and one of those inputs has gotten noticeably worse recently.

The click-through rate problem got worse

Search volume being approximate is a known and survivable problem. It is an annual average, it misses seasonality, and tools disagree with each other. You can work around that.

CTR is the input I would worry about. Depending on which study you read, position one gets somewhere between roughly 19% and 40% of clicks. That is not a rounding difference. If you build a forecast on the top of that range and reality lands at the bottom, you have overpromised by half before anyone has written a word of content.

The reason for the spread is that a ranking position is no longer one thing. Seer Interactive’s 2026 AI Overview CTR study, built on 53 brands and 2.43 billion organic impressions, found that informational queries showed an AI Overview about 36% of the time while transactional queries showed one about 5% of the time. Across a full year of informational queries, average organic CTR was 3.35% when no AI Overview appeared, 2.07% when one appeared and the brand was cited inside it, and 0.94% when one appeared and the brand was not. Same position on the page. Roughly three and a half times the difference in clicks.

So an honest keyword forecast now needs to know what the search engine results pages look like, not just where you sit on them, because changes in search engine results affect CTR assumptions. Comparison queries in that dataset triggered an AI Overview 95.4% of the time. If your content plan leans on “X vs Y” pages, that is the number to plan around, especially when shifts in user behavior on the results page change click expectations.

I want to be fair about the source. Seer is an agency, and agencies that publish search research have an interest in search staying interesting. They also published their own caveats, including that a single account made up 47% of the impressions in one segment, and that they cannot claim causation. That is more disclosure than most vendor studies bother with.

A forecast that missed, from people who forecast for a living

Here is the part of that study I have not stopped thinking about.

Seer built regression models, applying mathematical models to historical data, on twelve months of 2025 data and projected forward into the first quarter of 2026. That kind of statistical forecasting uses historical data to predict future traffic trends. The model said organic CTR would keep declining. In January and February it did the opposite, and every informational and transactional organic segment beat the projection. Their own write-up puts the miss at 0.6 to 2.6 percentage points across every segment.

The same model predicted paid CTR to within 0.09 percentage points.

Sit with that for a second. A serious research team, billions of impressions, a reasonable statistical method, and the forecast was near enough to perfect on paid while being directionally wrong on organic inside of two months. Not because anyone was sloppy. Because paid search is a system with stable rules and organic search currently is not.

That is the most useful thing I have read about forecasting in a while, and it is not really about math. It is about which channel you are allowed to be confident about. If you are presenting both to a leadership team, say out loud that your paid numbers are tighter than your organic ones. Being upfront about which of your numbers is soft buys more credibility than a confident number ever will.

How to forecast SEO growth without lying to anyone

Give three numbers instead of one

Build a conservative case where you win the low-competition terms and nothing else, a moderate case, and an optimistic case that assumes the content lands and you earn links. Ahrefs’ own forecasting output does a version of this, showing a shaded band rather than a line, with roughly an 80% probability that the real result falls inside it. Keyword-driven projections should start with a validated list of target keywords. A single number implies a precision that nobody actually has.

Anchor the conservative case in something real. This is where low search volume keywords earn their place in a forecast, because a term at difficulty four with 90 searches a month is a thing you can commit to. Keyword analysis helps identify high-value keywords before assigning them to forecast scenarios. A head term at difficulty 60 is a hope. Put the hopes in the upside column and label them as hopes. Check macro-trend validation and search trends before treating keyword growth as dependable.

Separate the leading indicators from the revenue

Revenue is a lagging metric, and it is the one everyone stares at in month three, which is exactly when it has nothing to show. Impressions, indexed pages, and movement on a tracked list of buyer-intent terms will tell you whether the work is landing long before the money arrives. The SEO KPIs you agree to watch in month two matter more than the revenue line you projected for month twelve, especially if you want to tie the work to real business outcomes. Used well, traffic projections help allocate resources and prioritize SEO efforts against other marketing work.

Then put a revision date on the document. Plan on quarterly reforecasting to adapt to changing conditions. A forecast that gets updated when real data arrives is a plan. A forecast that never moves is a sales prop that quietly turns into evidence against you.

On tools and templates

There is no shortage of seo forecasting tools and broader seo tools options, from free spreadsheets to enterprise platforms with a scenario modeller built in. They mostly run the same arithmetic. What separates them is whether they let you set your own CTR curve and pull from multiple data sources instead of relying on one dataset.

That is the feature I would actually check for, because a generic curve is the single thing most likely to make your forecast wrong. Ahrefs builds a custom curve from your Search Console data for exactly this reason. If you would rather build your own SEO forecasting template, that is fine, and honestly a spreadsheet you understand beats a tool you do not. Pull at least 12 to 16 months of first-party data from Google Search Console as historical traffic data, calculate your own CTR by position, and use that instead, since a longer window helps surface seasonal trends. Third-party data from seo forecasting software like Semrush can add estimated traffic based on crawling data and help you benchmark competitors. Your site’s curve is the only one that describes your site, but combining first-party and third-party data improves seo forecasting accuracy and supports more accurate SEO projections.

What actually determines whether a forecast holds

The forecasts that survive are not the accurate ones. They are the ones where everyone agreed in advance on what would count as on track.

That is why we tie our own guarantee to forecasted performance benchmarks agreed before the work starts, rather than to a position in search results. It is a narrower promise than a lot of agencies make and it is one I can keep, because the benchmark is a number we defined together instead of a ranking Google decides. I have an obvious commercial interest in you finding that framing reasonable, so weigh it accordingly. It is also the reason I can never sensibly guarantee rankings.

Forecasting and measurement are the same conversation from opposite ends. Before the work, you are estimating future performance. Afterwards, you are working out how to calculate SEO ROI on what actually happened. Effective seo forecasting combines historical data with keyword forecasting rather than leaning on one method alone. In practice, keyword research and analysis are stronger when they use both first-party and third-party keyword data. If those two exercises use different definitions of success, somebody is going to feel misled in month seven, and they will be right.

Build the range. Write down the assumptions. Check seasonality, market conditions, and competitor benchmarking to identify organic competitors and compare the model with actual traffic before judging it. Book the date to revisit it. That is most of the job.