Draft: hidden from Google until Dan signs it off.
Field notePsychology · 4 min read · 27 October 2026
The search before
the search.
Most buyers don't start by searching for a business. They start with the problem: 'why is my power bill so high', not 'solar installer Gosford'. Google's own research calls the stretch between the problem and the purchase the messy middle, a loop of exploring and evaluating. The business that answers the problem search has a head start when the buyer gets to the category search.
Three searches,
not one.
Think of it as a ladder. The buyer starts with a problem search, moves to a category search once they know what fixes it, then a shortlist search once they've got names. Most marketing budgets sit on the middle rung only.
The table below shows the ladder in five industries. The left column is where the decision usually starts. It's also where there's typically the least competition, because most businesses never write for it.
| Problem search | Category search | Shortlist search |
|---|---|---|
| why is my power bill so high in winter | heat pump hot water Central Coast | [brand] heat pump reviews |
| knee hurts after running downhill | sports physio near me | [clinic] physio Terrigal |
| how much tax do sole traders pay | accountant for tradies | [firm] vs [firm] accountant |
| why do people add to cart and not buy | abandoned cart email software | [tool] pricing |
| how to get more donations without events | Google Ad Grants agency | [agency] ad grants reviews |
Most marketing budgets sit on the middle rung only.
Why the first answer
sticks.
Google's messy middle study (Rennie and Protheroe, 2020) found people loop between exploring and evaluating, and that a handful of biases shape who they pick. The six they tested include authority bias, social proof and category heuristics.
Two of those matter most here. Authority: the business that explained the problem clearly looks like the expert by the time the buyer is comparing. Category heuristics: if your answer taught them what to look for ('check the COP rating', 'ask if they do a gait assessment'), they now judge every competitor by your checklist.
That's the real prize. Not the click on the problem search, but writing the buyer's shopping list before anyone else does.
Not the click on the problem search, but writing the buyer's shopping list before anyone else does.
AI search makes
this bigger.
Problem searches are long and specific, which is exactly how people use AI search. Google says searches in AI Mode are on average three times longer than traditional searches. Longer questions describe problems, not product names.
So a page that answers 'why is my power bill so high in winter' properly is now doing two jobs: ranking for the search, and being one of the sources an AI answer draws on. See how ChatGPT picks who to recommend for the second part.
Find your own
problem searches.
You've probably got some already, hiding in your data. Three places to look:
- Google Search Console, Performance report, filtered to queries containing 'why', 'how', 'can', 'should' or 'is it'. Note that Search Console leaves out rare and anonymised queries, so treat it as a sample, not the full list.
- Your Google Ads search terms report: the question-style terms you're paying for by accident are a map of the problems people have.
- Your inbox and phone. The first sentence of an enquiry is usually the problem search, word for word.
Your turn
Someone types 'knee hurts after running downhill'. Which rung of the ladder is that?
They haven't decided what fixes it yet, so they're not looking for a physio. Answer that question well, and you're in the running before they start looking.
In my experience
What I see most often in audits: a keyword list that starts at the category term and a content plan that starts at the product. The business is competing on the one rung everyone else is on, and nobody has written the page that answers what the buyer typed first. The fix is rarely more budget. It's usually one or two good pages a month aimed at the problem, written by someone who actually knows the answer.
What to do
this week.
Write the ladder
For your main service, write one problem, one category and one shortlist search. Use the table above as the pattern.
Check your data
Filter Search Console for 'why' and 'how' queries. List the ten problem searches you already show for.
Ask the front desk
Ask whoever answers the phone for the first thing people say. That's your problem search list.
Answer one properly
Pick the most common problem and write a page that answers it, including what to look for in a provider. That's your checklist in the buyer's head.
Link it up
Link the problem page to your service page, so the buyer can climb the ladder without going back to Google.
Questions
people ask.
What is the messy middle?
Google's name for the stretch between a trigger and a purchase, where people loop between exploring and evaluating options. It comes from a 2020 study by Google's UK team.
Are problem searches worth targeting if they rarely convert straight away?
Usually yes, for content. They tend to be cheaper to rank for and they shape who the buyer trusts later. For paid ads, test carefully: many problem searches are research, not buying.
How do I find problem searches for my business?
Start with Search Console queries containing 'why' or 'how', your Google Ads search terms report, and the first line of real enquiries. Then check the 'People also ask' box on Google for each one.
Do problem searches matter for AI search?
More than before. Google says AI Mode searches are about three times longer than traditional ones, and longer questions usually describe a problem.
What changed
- First published. Every source checked on the day.
Keep going

Want me to check
yours?
A free audit from me: the 3 biggest fixes, ranked, and a 20-minute call.
