The newsletterOnce or twice a week. 1 hot take, 2 tool suggestions, 3 best practices.
No. 011
Your CRM is always going to suck.
At the start of every engagement, I tell the team the same thing. Your CRM sucks right now, I’m going to fix it, and it’s still going to suck when I’m done. Someone actually repeated it back to me the other day, so clearly it’s something people need to hear so that it clicks.
It came up again on a coaching call last week and this is pretty much how I said it:
It sucks right now. I’m going to fix it. It’s going to suck when I fix it too. ... And then something about the business is going to change, or we’re going to learn something about the business that we didn’t know, and then we’re going to have to change it again.
Where I see this go wrong is when someone in RevOps or a CRO decides what the problem is and what the fix is, and then the whole team works on that one fix for three months. That’s too narrow and the time frame’s too long. There are a lot of small problems in a company at any given time, right? So I’d rather fix a few of them every week.
Also, this goes for stuff you built yourself. If you’re the one who set up your CRM, you have to keep checking whether your own builds are even working anymore, and you have to be OK breaking them. I do it to myself all the time.
I’ll do it three weeks later with a client. I’ll just, I’ll build something. I’ll be like this, this, Who the fuck did this? ... This was me? Okay, cool.
Two I’d set up this week. One I’d skip.
Exa, TinyFish and Firecrawl, so your AI can read the web
These are the tools I connect to Claude when I need it to search for something or read a web page. They all do sort of the same job, so I have all three connected and use whichever one fits.
Exa is what I use to find things. When I searched for people posting about their first sales hire, it came back with the actual announcement posts with dates on them, and the other search tools gave me a bunch of advice posts from consultants. TinyFish is for reading a specific page, and it got through a LinkedIn post with all the comments, which Firecrawl refuses to do. Firecrawl is what I keep around for crawling a whole site or pulling a list of every page on it.
Free tiers as of 10 November: Exa gives $10 in credits a month, Firecrawl 1,000 credits a month, and TinyFish’s Search and Fetch APIs are free (its agent and browser tools cost credits).
Claude’s Chrome extension, for testing what you built
A lot of the work I do results in pushing updates to customers’ CRMs. But someone still has to log in as a user and click through the build to make sure it works as intended. I was paying someone to do that QA, and every round trip meant things would slow down.
In the last couple of months the Claude Chrome extension started running a lot longer in the browser and using fewer tokens doing it (at least, that’s been my experience). Now my new setup is that one agent builds, a second one tests in the browser and sends feedback to the first, and they go back and forth until it works. It also gets through tools that don’t have an MCP (a way for AI to connect straight to another app), because it uses the screen the way a person would.
Now I just have one agent with the build, with sending stuff to HubSpot, and then the other agent live testing, sending feedback back, and it’s just going to loop. And I haven’t touched shit.
On Pro it runs Haiku 4.5; Max, Team and Enterprise let you pick the model.
HubSpot Enterprise
I keep running into companies under $1M ARR on HubSpot Enterprise. When I ask how they ended up there, it’s usually that someone in HubSpot’s partner program sold it to them.
At that size you need the CRM to hold your data. Connect Claude or ChatGPT to it, layer a few cheaper tools on top, and you don’t need half of what the higher tiers sell you. I wrote up the tier-by-tier version here.
A lot of times they’re like sub 1 million ARR, and they’re like getting sold on HubSpot Enterprise. Whoa, hold on. ... Like, what do you need this for? Like, what did the sales rep tell you that you needed for this?
Replacing it with
Three things I’d do this week
Tell your reps they’ll be slower first.
When you change how a team runs, they’re going to be slower at the new way for a while before it saves them any time. Say that before you get started and adjust expectations. If you change the process and leave every goal and quota where it was, the team will fight you on it.
You’re going to be slower and worse at doing it this way for some number of reps, right? And then out on the other side, you’re going to save this much time.
Write down your stage definitions before AE number two.
During founder-led sales you know what every stage means. You explain it to your first hire and it mostly holds. Every hire after that learns it from someone who learned it from someone else, and by the tenth one the definition has drifted from what you meant. Write down what has to be true for a deal to sit in each stage while you still remember. (I walked through what that list looks like in 009.)
And then you get the game of telephone. ... Where it just slowly deteriorates over time to the point where the founder, when they did founder-led sales, goes and asks the AE number 10, what’s the definition of this stage? ... And they repeat it back to them. Complete nonsense compared to what happened then.
Find where your CRM number and your finance number split.
A lot of companies just accept that the revenue number in the CRM and the one finance has are never going to match, and then they keep chasing down why. Sales says we sold this much, finance says we only collected about 80% of it, and people spend their time hunting for the other 20%. Find the exact spot where the two numbers split, which is usually a handoff between two systems, and fix it there.
It’s a distraction, right? It distracts from making decisions about actually important things because they’re just trying to track down the fact that the money’s wrong, right?
Thanks for reading. If you’re about to start a big CRM cleanup project, reply and tell me what’s on the list. I’ll tell you which pieces I’d do this week.
Kevin
The newsletter
This one went out in November 2026. Get the next one first.
Once or twice a week. Quotes from the calls, the tools I would buy, and what I would do this week.
No spam. Unsubscribe in one click.
Generate revenue. Subtract everything else.
Leads slip through cracks, reporting is a mess, and growth feels stuck. You’re not missing tools. You’re missing the system that ties them together.