I’ve been the HubSpot administrator inside a startup, partnered with marketing firms to stand up CRM ops for their clients, and built more than a dozen CRMs as a consultant — mostly HubSpot and Pipedrive, with budget-friendly setups in tools like Notion and Mailchimp when that was the right fit.

Across all of that, the pattern is almost always the same.

When someone tells me their CRM is broken, the platform almost never turns out to be the actual problem. It’s that nobody’s managing it, the data’s dirty, the structure has duplicate or missing fields or fields nobody fills in, and the team isn’t following any kind of process. And because the data’s bad, you can’t trigger marketing on top of it — which is honestly the whole reason most people bought the CRM in the first place.

This guide is for founders and operators who want a CRM that actually works. I’ll walk through how to scope what your tech stack needs to do, how to pick a platform when stage isn’t really the driver, and how to fix the boring stuff underneath that makes everything else possible.

Step one: scope the whole tech stack, not just the CRM

Most CRM projects start in the wrong place. The founder decides they need a CRM, picks one, and then tries to bend their business around the tool.

The right starting point is broader. Before you pick a CRM, you need to understand what technology needs to do for your business — across sales, marketing, and operations. The CRM is one piece of that stack. Your marketing platform might be a separate piece, or it might be bundled into the CRM (HubSpot’s Marketing Hub does both). Your operations layer — proposals, project management, invoicing, fulfillment — sometimes lives in the CRM and sometimes lives in adjacent tools that need to talk to it.

I see this wrong-starting-point pattern in early-stage businesses constantly. Someone signs up for HubSpot’s free tier to manage contacts, then realizes they need marketing emails so they bolt on Mailchimp, then needs proposal templates so they add Dubsado, and now the data is in three places and none of it talks. The CRM didn’t really fail — the scoping did. Nobody mapped out what the whole system needed to do before tools started getting bought.

I officially do marketing ops, but a lot of my early conversations with a new client end up being broader than that — because if the CRM has to run more of the business than its title suggests, somebody has to think about the whole stack. The CRM decision ends up being downstream of that thinking, not the other way around.

How to pick a CRM (when “stage” isn’t really the driver)

Most CRM advice tells you to pick by stage — pre-revenue, early traction, growth. Stage is a useful rough frame, but it’s not what actually drives the decision once you’re in the room.

In my experience the real drivers are four:

Budget. What can you sustainably spend on the stack — not just this year but next? CRM seats stack up faster than people expect, and once you add a marketing license on top, the total tech spend is usually 2–3x what the initial CRM bill looks like.

Growth expectations. How fast are you adding contacts, deals, and team members over the next twelve months? A tool that fits a team of three is painful when the team is twelve. Don’t overbuy for year two, but don’t undersize for year one either.

Seat count. Most CRMs price by user. Three seats versus fifteen seats is the difference between a free tier and a four-figure monthly bill. Forecast seats honestly before you commit.

Sales team structure. This one matters more than most founders realize. If you have a dedicated sales team — even one or two reps — you need a real sales hub with pipeline visibility, deal stages, and reporting they’ll actually use. If sales is the founder and one part-time closer, you can run leaner. The shape of the sales team should shape the shape of the tool.

Stage is a useful frame, but these four are what actually drive the conversation when I’m in the room with a client.

The platforms I scope for

These are the tools I’ve built and rebuilt enough times to have a real opinion on.

HubSpot is my default for B2B businesses serious about marketing. The free tier is genuinely useful and the paid tiers scale into real marketing automation. If you’re buying it primarily for sales, you’ll overpay — the Marketing Hub bundle is where it earns its cost.

Pipedrive is the cleanest sales-pipeline CRM on the market. If your team’s main job is moving deals through stages and you don’t need heavy marketing automation, this is what I’d recommend. It does one thing well and doesn’t try to be everything.

Notion as a CRM works in a narrow band. If you have an ops-minded contact who can configure Zaps and you don’t need a real pipeline, it can be a viable budget option — I’ll put the documentation together if that’s what you’re doing. But I tend to shy away from this when it’s not meant to be a CRM. The work to make Notion behave like one is more technical than marketing, and it scales badly.

Mailchimp as a CRM is for when the budget is tight and you just need a place to keep contact information clean. It’s clunky. You don’t get a pipeline. But you get contact records and you can email them, and for a single-founder business early on, that’s sometimes enough.

Salesforce is its own world. If your business actually needs Salesforce, you need a Salesforce specialist — not a marketing ops consultant — and I’m honest about that when it comes up. Honestly, hiring the right specialist usually costs less than asking the wrong one to fake it.

The real problem is rarely the platform

Here’s the part that took me years of CRM work to internalize.

When a client comes to me saying their CRM is a mess, the platform almost never turns out to be the issue. The issue is almost always one of three things — and usually all three at once.

Nobody is managing it. There’s no owner — no person whose job it is to maintain the system, train the team on it, enforce the rules, and clean the data on a cycle. Without an owner, every team starts using the CRM differently within six months, and within a year it’s basically a graveyard.

The data isn’t clean. Contacts have inconsistent capitalization, companies are duplicated three different ways, source fields are filled in for some records and blank for others, old leads from two years ago sit right next to active customers with no flag to distinguish them. None of these are catastrophes on their own — they just compound until you can’t trust any report you try to run.

The structure has problems. Fields that should exist don’t, and fields that do exist are duplicates of each other (“Lead Source” and “Source” and “How Did You Hear About Us” all somehow coexisting). Custom properties got added by three different people who didn’t talk to each other. Pipeline stages don’t match how the team actually sells.

None of this is exciting work, but it’s the work that actually matters.

Why this matters: you can’t trigger marketing on top of bad data

Here’s the part most founders miss until it’s in their face.

A CRM isn’t just a database — it’s the foundation for everything you actually want to do downstream of it. Segmented email, lifecycle automation, lead scoring, attribution reporting, retargeting audiences. None of that stuff works without clean data and clean structure underneath.

If your “engaged leads” segment is built off a custom field that only gets filled in half the time, your nurture sequence is going to a random subset of people. If your “trial signed up” trigger fires off a property that has three different naming conventions across records, the automation will fire sometimes and not others. If your reporting layer is built off duplicates and dirty source data, every decision you make off it is shakier than it looks.

This is why I care so much about the boring stuff. You can’t build the fun marketing layer — automation, lifecycle, scoring — on top of a CRM nobody is managing. Process and hygiene aren’t really optional; they’re the foundation everything else has to sit on.

What “managed” actually looks like

When I build or fix a CRM for a client, “managed” comes down to four things.

One owner. Someone whose job description includes the CRM. They don’t have to be full-time on it, but they’re accountable for data quality, structural decisions, and team adoption.

A documented process. What goes into the CRM, when, by whom, and in what format. Pipeline stage changes. Source field population. Contact creation. Activity logging. Written down somewhere the team can reference.

A regular hygiene cycle. Quarterly or monthly, depending on volume. Someone runs duplicate cleanup, fills in missing required fields, archives or merges stale records, and audits whether the structure still matches how the business actually operates.

A reporting layer that gets used. A weekly pipeline report. A monthly source-attribution view. Something the team sees regularly that makes the data visible. Adoption follows visibility — if the CRM produces reports the team relies on, the team will keep the data clean enough for the reports to work.

That’s the minimum. None of it is glamorous, but it all compounds over time.

When to bring in help

If you’re reading this and you feel some combination of “I know I need this” and “I do not want to think about this” — that’s the signal.

CRM build-outs and cleanups are one of the highest-leverage projects you can outsource. A well-architected, properly managed CRM compounds for years. A badly-built or unmanaged one creates technical debt that takes longer to unwind than it took to install. Most CRM projects I run for clients take 4–6 weeks: scoping the tech stack, configuring or rebuilding the CRM, cleaning the data, documenting the process, and training the team on adoption.

If you’re trying to do that on top of running the business, the project will stall. And a stalled CRM project is worse than no project, because you’ll spend six months convinced you’re “almost there” while the data quietly continues to rot.

The point

A CRM isn’t really a tool you buy — it’s a system you operate.

The platform matters less than the structure, the structure matters less than the management, and the management only ever happens if someone owns it, the process is documented somewhere, and the data gets cleaned on a regular cycle.

Get those three right and your CRM can finally do what you actually bought it for: give you the data foundation to run marketing on top of. Skip them and no platform on earth is going to save you.

---

Need help getting your CRM out of graveyard mode? I scope tech stacks, build CRMs, and fix the ones that have stopped working — mostly in HubSpot and Pipedrive. See the Marketing Ops & CRM service or book a Half-Day Intensive to audit what you have.

Free resource: Grab my Notion Marketing Planning Template for a starter framework you can adapt.