New format
Website traffic forecast: how to estimate SEO potential
Owners want a number: “how many visits will SEO bring?” There’s no exact answer: seasonality, competition, and demand shift. But you can estimate approximate potential — enough not to plan on maybe.
Below: the estimation logic — keyword set → frequency → expected positions → CTR by SERP place → sum. Site prep for promotion and page-one visibility for the keyword set are different horizons: rankings are planned over 2–6 months after work starts.
Why estimate if accuracy isn’t perfect
An approximate estimate answers practical questions: what traffic order is realistic, where to put content and budget, which clusters to pull first. Without a model it’s easy to promise “a million” or, conversely, underrate a niche.
Unpredictable factors (updates, season, new competitors) remain. So a forecast is a decision cue — not a contractual visit guarantee.
Four pillars of the model
A working scheme comes down to four blocks: relevant queries, demand volume for each, expected position (or a position corridor), and CTR for those positions. Then — the sum of expected clicks across the keyword set.
On large sites you may calculate priority sections and extrapolate — accuracy falls, order of magnitude usually holds.
Inputs:
- keyword set / clusters
- frequency (no double counting)
- achievable position estimate
- CTR by SERP place
Step-by-step calculation
Gather queries from demand stats, not “as the owner feels.” Planners and keyword tools give frequency order of magnitude but poorly catch season and sometimes merge close phrasings — don’t sum duplicates.
Estimate competition: who’s on page one now, how heavy the landings are. That decides whether to assume positions 1–3, 1–5, or 1–10 in scenarios. CTR doesn’t fall linearly: the gap between 2nd and 10th is multiplicative, not “a bit less.”
For each cluster: frequency × expected position CTR (or a corridor average) = expected visits. Sum across the keyword set. You get a ceiling under the chosen position scenario — not tomorrow’s analytics fact.
Step order:
- keyword set and clusters
- clean frequency
- position scenario
- CTR and click sum
- check against fact (if the site is live)
How to read the result
The model’s output is growth points: which clusters drive most volume, where positions are already close and you need CTR/snippet work, where the keyword set is thin. Sometimes it’s better to strengthen 20 phrases near positions 8–10 than spray across hundreds of zeros.
Check the forecast against actual traffic and visibility. A big gap is a signal: wrong positions in the model, cannibalization, tech issues, or inflated frequency. Then you fix the model and priorities — not “more budget at random.”
After the calculation, lock:
- cautious and base scenarios
- priority clusters
- what blocks fact from catching the model
- work horizon separate from the visit figure
FAQ
Can I forecast traffic exactly?
No. There’s a model with assumptions. Use a range and scenarios (cautious / base), not one “guaranteed” figure.
How does an SEO forecast differ from Google Ads?
In Ads the planner leans on auction and bids. In SEO — on demand, competition, and achievable positions. Different models.
Is summing keyword frequencies enough?
No. You must account for duplicates and merged queries, click share by position, and that you won’t take page one across the whole set at once.
Why not sum “buy car” and “car buy”?
Planners often show the same demand pool. Adding them double-counts.
When should I forecast — before launch or on a live site?
Both. At strategy stage — order of magnitude; on a live site — check against fact and find growth points.
Is a forecast the same as time to page one?
No. A forecast is visit potential at certain positions. Time to build rankings for the keyword set is separate — usually planned 2–6 months.
One “guaranteed” SEO visit number — and a TOP-in-a-month plan on top?
We’ll model demand, positions, and CTR as a range — potential separate from the planned 2–6 month TOP horizon.
Discuss the task
