Finding the buyer's problems
How to find, verify and cull the problems a buyer has, before you decide what to offer.
What it can see: only this page.
The runbook itself
A step in choosing what to offer. It answers four questions: how to generate the buyer's problems, how to verify they exist, what mental state to ideate in, and how to cull them.
The mechanism rests on three public frameworks: the four forces of progress (from jobs to be done), COM-B (Michie, van Stralen and West, 2011) and the theory of constraints (Goldratt).
1. What a problem is
A purchase is a switch from the buyer's current way to a new one. A switch happens when the push of the current situation and the pull of the new way beat the anxiety about the new way and the habit of the old one, and the buyer has the capability and opportunity to act.
So a buyer's problem is anything that holds them in the current way:
- Anxiety about the new: it might not work, it might embarrass them, they've been burned before.
- Habit of the old: the current way is familiar and already paid for.
- Missing capability: they can't, or don't know how.
- Missing opportunity: no time, no money, someone else's permission.
- Weak push or pull: the current pain isn't sharp enough, or the new way isn't vivid enough.
The generic list ("no time, no budget, unclear ROI") fails because it names kinds, not forces at a moment in a real person's path.
2. The mental state
Hold the buyer's struggle at a moment on their timeline, not a type of person: "Darren on a Tuesday night after a shoot, with three inquiries unanswered". Keep in view who judges them for the choice; people weigh relative status in what they choose. Forget our solution until section 5.
3. Generate from real switches, weighted by source
Imagined buyers miss the hidden forces (anxiety, habit). Start from real people, and weigh each source by who is in it and whom they were performing for. Every source is a trace of a choice in its own setting.
| # | Source | Weight for "is it real and common?" | Best use |
|---|---|---|---|
| 3.1 | Search queries, by people and their agents (a ledger of the queries) | highest | How common a problem is, the buyer's own words, and anxiety and habit ("is X worth it", "X scam", "how to do X myself") |
| 3.2 | Behaviour traces: Kokumi drop-off, what competitors sell and at what price | high | Where buyers stall, what they already pay to fix |
| 3.3 | Support requests | medium: buyers only | Problems after the purchase |
| 3.4 | Tender and RFI requirements (CanadaBuys) | a different job: not buyer problems | What the authority counts as a status marker (see below) |
| 3.5 | Interviews (the journalist skill: one real occasion, no leading questions) | lowest for frequency | Rebuilding one real timeline: how the forces built up |
| 3.6 | Reviews and forum complaints | lowest | Leads only, to check against 3.1 |
Why search leads:
- It is an act, not a statement, made at the moment of struggle, just before a switch.
- It has no audience, so no one performs status in it. Reviews and forums are written to be seen.
- It samples the silent buyers. Complainers select themselves; the best customers are busy, can pay, and mostly speak only to a search box.
- It is counted at population scale, so it says how common a problem is.
- It includes people who did nothing, and they show the hidden forces.
- It is also how they find us. The query is the problem in the words they'll use to look for the answer: information scent, the cues a searcher follows toward an answer (Pirolli and Card).
Tenders and RFIs answer a different question. They are a signal of what a nation treats as status markers: the government is the authority on what counts as credible. Their requirements (certifications, clearances, accessibility reports, pre-qualified lists, references) tell you which proof raises a buyer's standing.
Rebuild timelines from three kinds of switcher: people who bought a competitor, people who did it themselves, and people who did nothing. At each step, ask which force in section 1 held them there.
4. Verify by what they did
A problem is real when it shows in behaviour, not words:
- Seal a prediction first (
crux). - Look for behaviour, weighted as in section 3: people search for it, paid to remove it, stalled at that step, or stayed put. A complaint alone is a lead, not evidence.
- Two independent sources. One quote is a candidate.
- Score the prediction. A miss is signal.
5. Cull to the binding force
Only one force decides the switch at a time; removing any other changes nothing yet.
- For each kind of buyer, name the one force whose removal would flip the switch. Ask: if only this were gone, would they buy?
- Build the offer to remove that force first. It is the "without" in "get [result] without [binding constraint]".
- Keep the next forces in line behind it, for after the first one moves.
- Cut problems with no behavioural evidence, and problems that cost us more to remove than they are worth to the buyer.
6. What the step hands on
A table, one row per force: the buyer, the moment on their timeline, the force (section 1), their own words, the behavioural evidence and its source, the sealed prediction and its score, and whether it is the binding force. Step 4.3 turns the binding force into the offer's main promise.
Use it
Hand it to your agent: give it this page's address, or paste the runbook into the chat, and ask it to follow the steps.