Work through a problem
When something is not working, the first idea that comes to mind is rarely tested. Sorting what you see from what you assume helps you describe the problem clearly and choose a small test before committing to a fix.
Separate symptoms, evidence and assumptions, write a specific problem statement, and run a small test before choosing a fix.
Sort what you know
A symptom is what you notice: few people complete the form. Evidence is what you can point to: the form tool shows that most visitors who open it leave on the payment page. An assumption is something believed without checking: the logo looks unprofessional and puts people off.
Listing these in three columns often changes the discussion. Assumptions are usually the largest column, and they are where fixes tend to come from, which is why untested ideas can feel so convincing.
See the idea
in a real situation.
A community swimming club notices that few people finish its online signup form. At the next meeting, the first suggestion is to commission a new logo.
Meeting notes: 'Hardly anyone finishes signup. Probably the logo looks dated. Let's get a new one designed.'
Symptoms: few completed signups. Evidence: in this fictional example, the form tool suggests visitors often leave at the payment page; two members ask why the price is not visible earlier. Assumptions: the logo is the cause; people are not serious about joining. Problem statement: visitors who start the form often leave before seeing the total cost. Options: new logo, clearer questions, price earlier. Small authorised test: show the existing total cost earlier in a practice or agreed test version. Compare completed/start proportions over comparable periods, consider traffic and audience differences, and record what the result does and does not show.
↗Sorting evidence from assumptions shows that the logo idea has nothing behind it yet, while the payment-page pattern does.
↗The problem statement describes where people leave without naming a cause, so several explanations remain open.
↗A small agreed test can provide evidence before a larger change. Comparing proportions and context is more informative than comparing raw completion counts from different amounts of traffic.
Write a specific problem statement
A problem statement describes who, where and what, without suggesting a cause. 'People don't trust our site' already contains a theory. 'Visitors who start the signup form often leave at the payment page, before they have seen the total cost' states only what the evidence shows.
The statement should be narrow enough to test. If it could be solved by a dozen unrelated actions, it is probably still too broad.
Compare options and test small
List possible next steps, including the logo, and ask of each: what would we expect to see if this were the cause? A new logo would help if people leave because of appearance, but they leave late in the process, after investing effort. Showing the price earlier would help if the payment page surprises them.
Choose a small, authorised test that helps distinguish the options. You might observe people using a practice copy of the form, or compare completion proportions over comparable periods in an agreed test. Calculate completed forms divided by forms started, and consider differences in traffic or audience before drawing a conclusion.
Make it useful
in your own context.
Choose a real problem at work or home where a fix has already been suggested, and examine it before acting.
- Write three short lists headed symptoms, evidence and assumptions, and place every statement about the problem under one of them.
- Write a problem statement in one sentence that says who, where and what, without naming a cause.
- List two or three possible fixes, choose the cheapest test that would tell them apart, and write down what result would make you change course.
A problem worth checking
Make the example your own. Use fictional or non-sensitive details while you practise.
Your practice stays on this page. Copy or download it to keep a useful next step; the website does not submit your notes.
Which of your assumptions felt most certain, and what would it take to check it?
Before you adopt the first fix, write the problem statement and design one small test that could prove the idea wrong.
Explain what the evidence supports.
Check the information behind your conclusion, and shape a small next proposal that another person can understand.