"Should we sort out the CRM first?" is a question I get a lot, and nine times out of ten the person asking has not yet sold anything to a stranger. They have a product, a feeling about it, and a list of software to buy. The CRM is on the list because it feels like a business thing to own, the way a logo does.
The answer at that stage is no. A business runs on two variables, traffic and offer, and a CRM is neither of them. It cannot make people find you and it cannot make what they find worth buying. Before the three checkpoints have been passed, a spreadsheet with a name, an email, a date and a note is a complete CRM, and it costs nothing and teaches you what your real one will need to hold.
The day the answer changes
There is a day when the question flips, and it is easy to name. It is the day the thing you count stops being a sale and starts being a step before the sale: a booked call, an accepted quote, a job in the diary. The moment that step is the number on your wall, everything between the click and that step is where your money leaks.
At a mortgage brokerage I did the marketing for, the advisers had real expertise and the ads produced real people who were about to buy a house. The offer was fine and the traffic was fine. What broke was the handover: a person clicks an ad, books a call, and twenty minutes later nobody can say which ad it was, the booking is on one adviser's calendar and in nobody's system, and the follow-up depends on who remembers. The brokerage paid for a scheduling seat per adviser to fix that, and the only thing the seats did that mattered was put the booking in the CRM with the ad that caused it.
The back end of an offer is four small things, not a platform: capture who the person is, let them book, write it down somewhere with its source, and follow up without relying on memory.
Three signs you are there
You cannot say which ad, post or referral produced last month's best customer. If the answer is a guess, the traffic lever is being pulled blind, and no amount of extra traffic fixes a leak you cannot see.
A lead waited more than a day because nobody owned it. One person being away should not cost a sale. If it does, the handover lives in someone's head.
You are paying per seat for a tool whose only job is passing a name from one box to another. The brokerage's scheduler did exactly that. The booking page itself was a commodity; the Google Workspace they already paid for included one.
Two of those and the back end is the next thing to work on, ahead of a rebrand, ahead of a new website, and ahead of more ads.
What to do, in order
The record comes before the software. Write down what the business actually decides about a lead, where they came from, what they asked for, who owns them and what happens next, and put that in whatever you already have. Then pick a simple system that holds those fields and nothing else; the question that sorts the tools is where a won deal goes next, and for most small businesses the honest answer at step one is a filing cabinet with a good API, which is what Capsule is.
Then the booking. Every Google account already has a free booking page, and the reason nobody uses it with a CRM is that it has no webhook, no redirect and no prefill. The way round that is to put a short form of your own in front of it, so the source is captured before the visitor picks a time, and to watch the calendar for the booking instead of waiting for the page to tell you. I have published the whole thing, with the code, as free appointment scheduling for Capsule CRM, and it runs on my own site.
Then the follow-up. One templated email when the booking lands, one task for the person who owns the lead, written by the system and not by whoever noticed. That is the whole back end for a business that sells through conversations, and none of it is new software.
What not to do
Do not buy the platform at step one. The all-in-one suites are built for step four, when you have a data layer and automation worth paying a meter for, and their pricing is designed to grow with the automation you build inside them. Buying one before you know what your business looks like means paying platform prices to find out.
And do not let the back end become the new overwhelm. The list of CRM features is as long as the list of marketing tactics, and the same people sell both. Capture, book, record, follow up. When those four run without anyone remembering to do them, go back to the two levers, because that is where the next customer comes from.