SEO forecasting: how to estimate traffic without overpromising

Export at least 12 to 16 months of clicks by month from Search Console. Fit a trend that accounts for seasonality, using a spreadsheet forecast function or a simple model, and project it forward. This answers 'where are we heading if we continue as we are?' It cannot model new pages or big changes, and it assumes the past pattern continues, which algorithm updates and AI features can break.

For planned pages, list each target keyword with monthly search volume, current position, a realistic target position, and expected CTR at that position. Expected clicks equal volume multiplied by CTR. Add a ramp: new pages on young domains take months to reach their target position, if they reach it. Multiply clicks by your organic conversion rate to estimate leads or signups.

Build a CTR curve from your own Performance data by grouping queries by average position and computing CTR per group, separately for branded and non-branded queries. Review the results page for each important keyword: AI Overviews, featured snippets, ads, and video results all lower organic CTR, especially for definitions and quick facts. Search volumes from tools are estimates and can differ widely between providers, so treat them as ranges.

Give three scenarios: low, where some pages never rank well; expected; and high, where most reach target. State the assumptions plainly, such as publishing cadence, target positions, and ramp time. Show the forecast against actuals each month and revise it. A forecast that is honest about uncertainty builds more trust than a confident number that misses.

There is no reliable volume data for AI assistant prompts, so AI visibility cannot be forecast like keywords. Instead, set goals for mention and citation rates on a fixed prompt set and track them. We forecast new pages on serionflow.com by indexing rate first, then impressions, before clicks, because on a young domain those move first and are easier to predict. SerionFlow does not produce traffic forecasts; this method works with a spreadsheet.