Think of the question you get asked the most. The one you could answer in your sleep. You've probably answered it a hundred times. Now try to show me those hundred times. Most people can't find one.
That's the gap. Your customers tell you what they want every week, in their own words, and almost all of it disappears the moment you hit reply. This is the system that keeps it: one place every question lands, so working out what to sell, write or fix takes two minutes instead of an afternoon of scrolling.
You answered it. You didn't keep it.
A question arrives in a DM. You answer it. It helps. Then it scrolls away under forty other messages.
A question comes up on a call. You answer it. The call ends. Nobody wrote it down.
Someone replies to an email with a question. You answer it. By Friday it's buried.
Every one of those was someone telling you exactly what they need, in their own words. That's the most valuable research you'll ever get, and it's free. Most of it ends where it arrived.
There are three stages to every question:
- The question. Someone tells you what they need. That's gold.
- The answer. You reply and you help them. That's the part everyone does.
- The gap. The question is gone, and the count goes with it. You helped one person once and the research went in the bin.
This system doesn't change how you answer. It adds thirty seconds before the answer, so the question stays.
Why nobody keeps them
Three reasons. You'll recognise at least one.
Memory. "I'll remember the big ones." You won't. You'll remember the loud ones, the weird ones, the ones that annoyed you. Not the frequent ones. And the question asked quietly ten times beats the one asked loudly once.
Inbox. "It's all in my DMs." Technically, yes. But a platform inbox is rented. You can't search it properly, you can't count it, and if the account goes, the questions go with it. That's not storage. It's a hallway.
Tool. "I need a CRM first." No. Software doesn't keep questions. A habit does.
All three trust somewhere else to keep it. The Question Bank means you keep it yourself.
The Question Bank: four parts
- The doors. Every place a question can reach you.
- The one place. Every door points to one place you own.
- The capture. Their words, where and when. Thirty seconds.
- The count. Once a week, tally the repeats.
None of it needs new software. Build it once, and every question after that lands somewhere you own.
Part 1: The doors
List every door a question walks through. Most people have five or six and only ever think about one: the inbox they check most.
Typical doors: DMs, email replies, calls, comments under your posts, in person, and a support address if you have one.
Now circle the doors that leave no record. Calls and in person are the usual culprits. A DM at least sits there. A call is gone the second you hang up. The annoying part is that the best questions often come through exactly those doors, because people say things out loud they'd never type.
The test: which of your doors leave no record? Those are the ones you're losing the most from.
Part 2: The one place
Every door points to one place you own. Not five inboxes. One.
What counts? One spreadsheet. Not a new one. The one you already open every day. One row per question. That's it.
What doesn't count? A platform inbox, a comment section, a group chat. And the one people get wrong: a support desk. A desk is brilliant at what it does, but it's a door with a log. Questions arrive, they get answered, they get closed. A closed ticket doesn't get counted. So the desk is a door like the others, and it points to the sheet.
Why the spreadsheet you already use? Because you'll actually open it. A new tool you have to remember to visit dies in a week or two.
Three words to check it against: searchable, sortable, yours. Sortable is the one that matters. Sort by question and the count does itself.
The test: if the platform closed tomorrow, would the questions still be there? If yes, you've got the one place. If no, it's a door, not a bank.
Part 3: The capture
Their words. Where. When.
Thirty seconds per question, and one habit makes it work: capture before you answer. Copy the question into the sheet word for word, then reply.
Word for word matters more than anything else here. Don't tidy it up. Don't turn "how do I get him to stop dragging me down the road?" into "lead training". The messy version is the research. That's how they say it, which means that's how they search for it, and that's the headline on the page you'll eventually sell from.
Then add the door it came through and the date. Done.
The test: could you lift it straight onto a sales page as the headline? If you tidied it, you can't.
Part 4: The count
Once a week, open the sheet and count the repeats.
Group the questions that are the same question in different words. "How do I stop him pulling?" and "He drags me down the road" are the same question. One stroke per repeat.
Anything that hits ten goes through the Ten Times Test: repeat, pain, wait. Three out of three is the door worth selling through.
This is where it pays off. Without the bank, running that test means scrolling through months of messages trying to remember. With it, the count takes about two minutes, because the questions are already sitting there in their own words.
The test: can you name your most-asked question without guessing? After a month, you'll know it to the number.
Three rules that keep it running
This is the bit people skip, and then it falls over in week two.
- Capture before you answer. Thirty seconds, then reply. If you answer first, you won't come back to capture it. Nobody does.
- Their words, never yours. The moment you paraphrase, you swap their language for yours, and their language is the valuable part.
- Count on a fixed day. Not when you remember. Put it in the diary like a client meeting. Count "when I get round to it" and you'll count twice and stop.
What it looks like in two real businesses
A local dog-walking business. Five doors: WhatsApp, phone calls, the doorstep at pick-up, Facebook comments and the booking form. One place: the same spreadsheet used every day. The first three rows don't need a count, because they come up all the time:
- "How much do I owe for the walks?" That's money. Every one is an owner who can't see their balance. Answer it before they ask and the question disappears.
- "Did he go to the toilet?" That's care. They're not checking up. They want to know their dog is OK. Same fix: tell them before they ask.
- "Do you look after dogs in your own home?" This is the one. That isn't a support question. It's someone asking to buy something. Count how often it comes up and you'll know whether there's a service there.
Three questions, three different jobs. Without the sheet, all three are just messages answered and forgotten.
An online business. The doors are replies to a weekly email, posts inside a private community, DMs, comments and a support desk. Every door, desk included, lands in one spreadsheet. Then the top repeats go one step further: onto a public ideas board, where people vote for what gets built next and follow each idea from idea to release. The bank stops being private research and starts shaping what gets built, in public.
How to collect customer feedback without a survey
Most advice on finding out what customers want starts with a survey. The trouble is that a survey asks people to predict what they'd want, and what people say and what they actually ask for are two different lists.
The Question Bank flips it. You're not asking for feedback. You're keeping the feedback people already volunteer, every time they ask you something. It's unprompted, it's in their words, and it arrives at the moment they actually need help. That's a better signal than any form, and it costs thirty seconds a question.
If you do want to ask, ask something that produces a question rather than a rating: "What's the one thing you're stuck on right now?" Then bank the answer like any other.
Why it compounds
In the SOI™ Framework this is Asset Anchoring: lock the value into something you own, so every interaction builds something permanent instead of feeding someone else's platform.
A question in a DM is rented ground. Two things happen when someone asks you something: the reply and the record. The reply is fuel. It helps one person once, then it's spent. The record is the engine. Kept, counted and reused, it turns into the FAQ, the blog post, the next product. Anything you don't own can't compound.
Run the Litmus Test: if the funnel disappeared tomorrow, what would you still own? With a bank, the answer is every question your people ever asked you, in their own words.
Build it today
Fill in this sentence and put the day in your diary:
Every question lands in __________. I count them every __________.
The second blank is a day, not "when I remember". DMs don't count as the first blank, and neither does a support desk or an inbox. A door with a log is still a door.
You can only count what you kept. Every question you didn't keep is research you gave away. Build the bank today, and the first question tomorrow is the first deposit.
