A problem impact quantification SOP turns "this is annoying" into "this costs you €6,000 a year", and it takes one to two weeks. You split the problem into the hours it eats, the money it burns and the risk it carries, collect real numbers from customers and build a small model with low, likely and high values. Customers buy fixes for costs they can see. A clear euro figure gives your pitch, your price and your sales conversations something solid to stand on.

What is the problem impact quantification SOP and when should you use it?

Problem impact quantification is a cost model for one problem, seen from one customer's side. It uses a simple formula: frequency x time per occurrence x hourly cost, plus direct spend, plus expected losses (the chance of a bad outcome multiplied by its size). The method borrows from Douglas Hubbard's How to Measure Anything: estimate every uncertain number as a range you are 90% confident in, and narrow the range with small samples. Hubbard's Rule of Five says that with five random samples there is a 93.75% chance the true median of the group sits between the smallest and largest value. Five customers can already tell you a lot.

Use this SOP:

When you will sell on savings: time, money, fewer errors, less risk. Most B2B products fall here.
When customers agree the problem exists but you don't know if it is big enough to pay for.
Before you set a price or pitch investors, since both need the cost of the problem as an anchor.

Scope check: this SOP measures what the problem costs the customer. What they are willing to pay is a separate question, answered by the problem monetization assessment SOP. How many customers exist is covered by market sizing.

What do you need before you start?

A validated problem statement with a known frequency.
10 to 20 customers from one segment willing to share rough numbers, ideally from interviews or a pain point diary with minutes already logged.
A spreadsheet with low, likely and high columns for every input (template below).
Access to records where possible: calendars, invoices, support tickets, timesheets. Records beat memory every time.

How do you run it, step by step?

1. Split impact into categories

List every way the problem costs the customer: time (staff and founder hours), direct money (fees, tools, rework, penalties), risk (fines, lost clients, refunds, safety) and missed revenue (sales that never happen). Note emotional cost, such as stress or embarrassment, in words. It matters in sales but stays out of the money total.

2. Write the cost formula per occurrence

For each category write the formula in plain terms, such as "returns per month x minutes per return x hourly staff cost". Use a loaded hourly cost (salary plus taxes and overheads) for staff time, or the customer's own billing rate for freelancers.

3. Collect inputs as ranges

Ask each customer for a low, likely and high value for every input: "On a slow month, how many? A typical month? Your worst month?" Ask them to check a record while you are on the call. Mark each input as record or estimate.

4. Apply the Rule of Five early

After five customers, look at the spread of each input. If the smallest and largest values are close, your median is already trustworthy. If they are far apart, you either have two segments mixed together or a badly worded question.

5. Compute yearly cost per customer

Multiply out low, likely and high for each customer and annualise. If you want more than three numbers, a free tool such as Guesstimate runs a Monte Carlo simulation over your ranges and returns a distribution.

6. Segment by severity

Sort customers by likely yearly cost and look for natural tiers: light, medium, heavy. Describe each tier by a visible trait (size, volume, industry) so marketing can find them later.

7. Run the 10x test and write the model

Compare the likely yearly cost of each tier with your planned yearly price. This SOP uses a 10x rule: a tier passes if the problem costs at least ten times your price, which leaves room for doubt, switching effort and your margin. Write a one-page model with the formula, inputs, tiers and verdict.

Template you can copy

IMPACT MODEL (one row per customer)
Customer | Segment trait | Occurrences/month (L/M/H) | Minutes each (L/M/H) |
Hourly cost € | Time cost/month | Direct cost/month | Risk events/year x
cost per event | Missed revenue/month | Yearly total (L/M/H) | Source: record/estimate

FORMULAS
Time cost/month    = occurrences x minutes / 60 x hourly cost
Risk cost/year     = expected events per year x average cost per event
Yearly total       = 12 x (time + direct + missed revenue) + risk cost/year
10x test           = likely yearly total / planned yearly price  (pass if 10 or more)

CALL QUESTIONS
1. In a typical month, how many times does [problem] happen? Slow month? Worst month?
2. How long does each one take, start to finish? Who handles it?
3. What do you pay, outside of time, because of it?
4. In the last year, did it ever cost you a client, a fine or a refund? How much?
5. Can you check that in your [calendar / invoices / ticket system] now?

Worked example: small online shops handling returns

This is an illustrative case. A founder plans a returns tool for online fashion shops and interviews 14 shop owners, 9 of whom open their order systems during the call.

For the median shop: 120 returns a month, 9 minutes each, at a loaded staff cost of €22 an hour. That is 18 hours, or €396 a month in time. Misprinted return labels add about €40 a month, and two fraudulent returns a month cost about €35 each, adding €70. The likely total is €506 a month, or €6,072 a year. Across all 14 shops, likely yearly cost ranges from €2,400 to €14,000.

The tiers are clear: shops with under 60 returns a month sit below €3,000 a year; shops with 100+ returns a month sit above €5,500. Her planned price is €49 a month, €588 a year. The 10x test fails for small shops and passes for shops with 100+ returns. She narrows her target to those shops and hands the numbers to her problem monetization assessment.

What should you measure, and when do you stop?

Metric	Pass signal	Warning signal
Yearly cost per customer	Likely value for your target tier is clear and repeatable	Every customer gives a wildly different figure
Cost-to-price ratio	10x or more for the target tier	Under 3x for every tier
Inputs backed by records	Half or more of the inputs come from records	All inputs are memory estimates
Estimate range width	High value under 3x the low value for the target tier	High value 10x the low value

Stop when new customers no longer move the median of the target tier by more than 10%, which tends to happen between 10 and 20 customers. If no tier passes 3x, the problem is real but too cheap to build a business on at your price.

Which mistakes should you avoid?

Using salary instead of loaded cost. Staff cost the business more than their salary once taxes and overheads are in.
Counting stress as euros. Keep emotional cost in words. Inflated totals fall apart in the first sales call.
Single-point estimates. "About 10 minutes" hides the slow and worst months. Ask for ranges every time.
Averaging all customers together. Heavy and light users average into a number that describes nobody.
Treating the cost as your price. Customers will not pay €6,000 to save €6,000. The 10x rule leaves room for them to win.

How can AI speed this up?

An AI assistant can turn call notes into the impact spreadsheet, write the formulas, flag inputs that look inconsistent across customers and draft the one-page model. It can also suggest cost categories you missed for a given industry, such as chargebacks for e-commerce. Check every input against the source, since AI fills gaps with plausible numbers. Once the per-customer cost is solid, multiply it by the number of customers in your target tier with the free market size calculator, and put the result on the problem slide of your deck with the F/MS pitch deck tool.

Where does this SOP fit in your validation plan?

Run it after problem statement validation and a pain point diary or interviews that gave you frequencies and minutes. Pair it with problem monetization assessment. Later, the yearly cost feeds unit economics validation and financial model validation. All ten SOPs of this stage are on the Problem Validation page.

FAQ

How do you quantify the cost of a customer's problem?

Multiply how frequently it happens by the time each occurrence takes and the hourly cost, then add direct spend and expected losses. Collect each input as a range from 10 to 20 customers and annualise.

What is the 10x rule in pricing?

A rule of thumb used in this SOP: the yearly cost of the problem should be at least ten times your yearly price. It leaves room for customer doubt, switching effort and your own margin.

What is Hubbard's Rule of Five?

From Douglas Hubbard's How to Measure Anything: with five random samples from a group, there is a 93.75% chance the group's median lies between the smallest and largest sample.

Can I quantify emotional impact?

Describe it in words and quotes, and use it in marketing. Keep it out of the money total, because customers and investors discount numbers they can't check.

Problem Impact Quantification SOP