Ask a builder what your business needs and you'll be told to hit something with a hammer. Ask a marketer and you get a list: a second social platform, a rebrand, a podcast, a CRM, a content calendar, a new funnel, and a few tools the person recommending them happens to get paid for. Each item sounds reasonable on its own, and together they're how a small business ends up busy, tired and no further ahead than it was in January.
Simplify My Business makes the cut for you, on your own numbers. You run it inside ChatGPT or Claude, it reads whatever you've already connected or uploaded, and it hands back three things to do, a list of things not to do yet, and a date to check back in. It's the diagnostic on this site, run on your business instead of described in general, and it's free, with no sign-up and no email address asked for.
Not now is not never
Most advice about focus tells you to say no, and most owners can't, because half the list is genuinely good. A podcast might be worth doing one day. So might the rebrand. The trouble is the timing, and a flat no throws the idea away along with the distraction, which is why the list keeps growing back.
So every item the tool parks gets a trigger, the thing that has to be true before it comes back: "when the product page converts above 2%", "after the Christmas rush", "when you've got ten reviews". A trigger is something you can check on a Tuesday morning, and that is the difference between a Not Now list and a guilty conscience. The few things that are a plain waste go on a third list, Never, with one line saying why.
The Now list holds three things at most. Ask for a fourth and the tool asks which of the three it replaces.
What it does
It works through five steps, one short message at a time:
- Five questions. What you sell and what one sale earns you, who buys it, the one result you want in the next 90 days, the hours and money you can put in, and everything you're doing, planning or feeling guilty about not doing.
- Your data. Whatever it can see: the store, the ad accounts, a spreadsheet you upload. It reads and nothing more, so it never changes a campaign, a price or a product.
- Whether you can deliver. If a product will sell out in three weeks, buying more traffic for it buys refunds, so stopping the overselling comes before anything else.
- Which lever is broken, and where. Overall numbers can look fine while one market or one product line leaks, so it checks each one and names the place, "offer, in the UK" rather than just "offer". When traffic and offer are close, it picks offer, because more people walking past a weak offer is the same result at a higher cost.
- The lists. Now, Not Now with a trigger on each, and Never.
Every number in the answer says where it came from, and when the tool can't see something it says so instead of guessing. Sales are taken from your store or payment system, because the ad platforms each claim credit for the same sale.
Your data stays where it is
The tool doesn't connect to anything itself. It uses the connections you've already set up in your own ChatGPT or Claude account, or the files you give it, and the conversation stays in your account. Nothing reaches me, and it never asks for a password or a key.
If something useful isn't connected, it suggests two sources at most, the ones that would change the answer, with the steps for the assistant you're using and the export to upload if you'd rather not connect anything. Suggesting ten would just be another list. As of 4 November 2026, Shopify and Google Drive have official connectors on both ChatGPT and Claude, and Search Console, Google Analytics and Google Ads have no simple one on either, so for those three the tool asks for a CSV export instead.
What you get at the end
The plan fits on one page. Under it the tool gives you a card to share, because saying no to nine things is a decision, and a decision is easier to keep when somebody else knows you made it. Here is one for a made-up business, two people selling candles online and at weekend markets:
My next 90 days
The constraint: Offer
Now: 1. Price and delivery cost on every product page · 2. One Christmas bundle · 3. Ask the last 20 buyers why they bought
Not Now: 9 things. Never: 2 things.
Not Now is not never. Each one has a date or a trigger.
The card carries no sales figures unless you add them. The tool also writes a short note you can forward to a business partner, a co-founder or whoever has to agree with the plan, carrying the one decision that needs both of you, and it gives you a date thirty days out to come back and run it again against the same numbers.
Where it stops
It's a framework, not a consultant. It can tell you the offer is the problem and which page is leaking, but not what your customers are thinking, and the cheapest way to find that out is still to ask the people who didn't buy. It also can't create a business from nothing: if nobody wants what you sell, no list fixes that, and the test comes before the plan, which is where the 3-checkpoint validation plan starts.