In one line
CLEAR answers two questions: how much organic revenue is at risk from AI answers, and how much of it does Profound need to win back to pay for itself?
It doesn't assume what Profound will achieve. It works out the hurdle and lets the buyer judge whether it is realistic. A small hurdle is easier to believe than a big ROI claim.
The five letters
| Letter | What it is | Where it comes from |
| C Cost | Profound subscription, setup, and the internal time to run it | Profound quote, plus a standard allowance for internal time |
| L Loss | Organic revenue × the drop in click-through rate | Prospect: analytics and Google Search Console |
| E Efficiency | Hours a week spent on manual AI tracking and reporting, turned into cost. Optional, zero by default | Prospect: one discovery question |
| A Avoided spend | Agency or tool spend on AI visibility reporting that Profound replaces. Optional, zero by default | Prospect: they know what they pay |
| R Recovery | The share of the loss Profound must win back to break even | Calculated by the model |
The maths
Revenue at risk (L) = organic revenue × drop in click-through
Efficiency (E) = hours a week × 48 weeks × cost per hour × 50%
Recovery needed (R) = (Cost − Efficiency − Avoided spend) ÷ Revenue at risk
The result reads: "Profound pays for itself if it wins back X% of the $Y a year at risk." Add the order or deal value and margin, and it also says how many extra customers or orders that is: net cost ÷ (deal value × margin). Below it, a table shows the value, ROI, NPV and payback if Profound wins back 10%, 20% or 30%.
Where to find the two key numbers
- Organic revenue. Three ways in. Use the revenue figure from web analytics if they have it. Retail and travel teams can use organic visits × conversion rate × average order value. B2B teams can use organic leads × lead to customer rate × deal size.
- Drop in click-through rate. In Google Search Console, compare the last 3 months with the same 3 months last year, on non-brand searches. (Search Console only keeps 16 months of data.) When impressions hold steady but clicks fall, AI answers are taking those clicks. If click-through went from 3.0% to 2.4%, that is a 20% drop. The "Send this to the SEO team" button gives your champion a ready-made request.
Until the prospect confirms it, the model uses a published benchmark: Ahrefs found 34.5% fewer clicks on the top result where an AI Overview appears. That is multiplied by the share of the prospect's searches that show one (18% by default, from Pew Research). Both studies use US data, so treat the result as a starting point for APAC.
How it stays conservative
The model follows the same principles as Forrester's Total Economic Impact method.
- Benefits are reduced for risk: revenue by 20%, time and spend savings by 10%.
- Only half of the hours saved count, since not all freed-up time turns into output.
- All the costs are counted, not only the subscription: setup and the internal time to run it too.
- 3-year figures are discounted at 10% a year, so future value is worth less than value today.
- Anything that can't be measured today is listed but not counted, including revenue from AI assistants, AI shopping placement and brand accuracy. Profound measures AI-assistant revenue during the pilot.
- Revenue at risk uses current revenue, which slightly understates the true loss.
Reading the result
| Recovery needed | What it means |
| Under 5% | A low bar. The case is strong. Move to a pilot. |
| 5% to 15% | A reasonable bar. Use the pilot to prove it. |
| Over 15% | A high bar. Firm up the inputs, or lead with the break-even in customers. |
What is calculated and what is written
Every number is calculated by fixed formulas in the page, so the same inputs always give the same answer. The AI only writes the CMO and CFO summary around those numbers.