Baseline
What happens without the work?Start with a realistic no-work case for the same period.
Test your baseline method against the pastFrom Rankings to Ranges · James Yorke at brightonSEO
Build an SEO forecast you can explain.
For SEO teams who need to estimate what their work could deliver. Learn the six-step Range Model, try a fictional example, then build a range with visible assumptions and a clear review point.
Here for the talk? Take the slides, then try the worked example. Free resources, no sign-up required.
FICTIONAL SHOE SHOP / 1 JANUARY–31 DECEMBER 2027
What could the work add?
Three scenarios for additional sales above the no-work baseline. These are not statistical confidence intervals.
Low+100
additional sales
More cautious assumptionsBest-supported+140
additional sales
Strongest support todayUpside+181
additional sales
More favourable assumptionsHow the best-supported case adds up
420No-work sales
+140Additional sales
=560Combined sales
These assumptions support about 100–181 additional sales over the same 12 months, with 140 in the best-supported case. A target of 250 additional sales still has a 110-sale gap. Review the assumptions before treating this as a funding commitment.
The middle case has the strongest support today. It is not a promise. Delivery and new evidence can change the range.
Try one small change
With 100,000 additional impressions and 2.8% conversion, 5% CTR supports 140 sales. At 4.57% CTR, that falls to about 128. At £8,000 cost and £60 contribution per sale, practical break-even is 134 whole sales: a small change crosses the decision line.
Try the simple calculationTHE METHOD
Six questions. One useful decision.
Group
What behaves differently?Separate work where search intent or results behave differently.
Plan groups that need different assumptions →Evidence
What supports the assumptions?Use your own results where possible. Label judgement clearly.
Build an evidence-backed assumptionScenarios
What could happen?Make Low, Best-supported and Upside versions that each make sense.
Move one assumption in Sensitivity LabDelivery
When will the work actually land?Include launch dates, dependencies and the time results take to appear.
Test what happens if launch slipsReview
What happened, and what changes?Compare with actual results, keep the original and update what remains.
Compare expected with actual and update what comes nextAPPLY THE METHOD
Build your forecast in Range Room.
Try the shoe-shop example or use one comparable group of your own figures. Keep the assumptions, a scenario range, decision context and a review point in a summary you can retain.
Build a forecast in Range RoomNeed help with one part?
Optional detours: you do not need to complete every tool to build a forecast. Carry findings back manually unless the tool explicitly says what transfers.
TAKE SOMETHING AWAY
Slides and forecasting spreadsheets.
The brightonSEO talk slides
James Yorke — From Rankings to Ranges: How to Forecast SEO in 2026.
Full slide deck · PDF · 72 pages · 1.6 MB
Take the method into your own work
Start with the Basic spreadsheet for one group. Use Advanced for several groups and a fuller review log.
Look up a term in the Field Guide
Optional practice and process checks
Practise ten forecasting decisions
Advanced shoe-shop learning example — a separate, more detailed teaching model.