Most HubSpot setups at an early stage look the same.
Someone got access, watched a few onboarding videos, customized everything they could find, and built a CRM that felt thorough. Lots of fields. Lots of stages. Lots of properties tracking lots of things.
And six months later, nobody trusts the data.
Not because the tool failed. Because the setup was solving the wrong problem.
Here are the five fields I see misconfigured — or misunderstood — in nearly every early-stage SaaS company I audit.
This is the big one.
Most founders look at HubSpot's default pipeline stages, decide they don't match their sales motion, and build something custom. Twelve stages. Detailed definitions. A stage for every nuance of how they sell.
The problem isn't the customization. The problem is what those stages are actually tracking.
Pipeline stages should reflect buyer decisions — not seller activities. "Demo scheduled" is an activity. "Technical evaluation in progress" is an activity. Neither of those tells you where the buyer is in their decision.
When stages track what your team is doing instead of what your buyer has committed to, two things happen. Reps move deals forward based on effort, not momentum. And your pipeline starts reflecting optimism instead of reality.
The fix: each stage should represent something the buyer did or agreed to — not something your rep did. If you can move a deal to the next stage without the buyer doing anything, the stage definition is wrong.
These two fields exist for a reason, and conflating them is one of the most common structural mistakes I see.
Lifecycle stage is a contact-level field that tracks where someone is in their overall relationship with your company — subscriber, lead, MQL, SQL, opportunity, customer. It moves forward and rarely moves back.
Lead status is a working field for your sales team — it tracks what's happening with a contact right now. New, attempting to contact, connected, bad timing, disqualified.
When founders don't understand the distinction, one of two things happens: they use only one field for both jobs (which breaks your funnel reporting), or they create a dozen custom lifecycle stages because the defaults "don't match our business."
They almost always match your business. What doesn't match is the workflow built around them.
Before you add a custom lifecycle stage, ask whether you actually need a new stage or just a new lead status value. Ninety percent of the time, it's the latter.
Close date is the most optimistic field in any CRM.
At seed stage, close dates are almost always set based on when the rep wants the deal to close — not when the buyer has indicated they'll make a decision. They get pushed forward every week. They've been in "closing this month" for three months. And because nobody wants to move them backward, they just drift.
The result is a pipeline that looks fuller than it is and a forecast that's wrong in the same direction every quarter.
Close date only becomes a useful forecasting input when it reflects a real signal from the buyer — a stated decision timeline, a procurement process, a board meeting they're waiting on. Without that, it's a wish field.
If your average close date has been pushed more than twice on the same deal, that deal needs a status conversation — not another date change.
This is where I see the most creative field architecture — and the most useless reporting.
Founders want to know where their revenue is coming from. That's the right instinct. So they build out a source field with fifteen options: LinkedIn outbound, LinkedIn inbound, conference, referral from investor, referral from customer, cold email, warm intro, website organic, website paid, and so on.
Then they don't fill it in consistently. Or reps interpret the options differently. Or the most common source is "other" because nothing quite fits.
And the report they pull looks detailed but doesn't actually tell them what's driving revenue.
The better approach is to start with fewer, cleaner categories and fill them in with discipline before adding nuance. Five sources you trust are worth more than fifteen sources that are half-populated.
More importantly: track the source that led to revenue, not just the source that generated a contact. Those are often different. A founder referral that came in through the website shouldn't be credited to organic search.
This one is less about configuration and more about honesty.
At seed stage, deal amounts are frequently entered as the hoped-for contract value rather than the expected contract value. Multi-year deals get entered at full ACV. Expansion potential gets baked into the initial number. Best-case pricing gets used before any negotiation.
The result is a pipeline that looks healthy on paper and a forecast that consistently disappoints.
Deal amount should reflect what you actually expect to collect in the current contract term, at the most likely price point, before you've won. Not what the deal could be worth. Not what you're going to propose. What you'd bet on.
If your average deal size in the CRM is consistently higher than your actual average deal size at close, the field isn't being used correctly — and every forecast built on it is starting from a wrong assumption.
Every one of these mistakes comes from the same place: tracking what's flattering instead of what's true.
Stages that reflect activity make the pipeline look busy. Optimistic close dates make the forecast look full. Inflated deal amounts make the quarter look achievable. Overly complex source fields look sophisticated without producing insight.
The CRM that tells you what you want to hear isn't an operating system. It's a story you're telling yourself.
The founders who get the most out of HubSpot aren't the ones who customized it the most. They're the ones who configured it to reflect reality — even when reality is uncomfortable — and then actually used what it told them.
That's what makes forecasting reliable. That's what makes pipeline reviews useful. And that's what turns your CRM from a logging tool into something that actually runs your business.
Photo by GuerrillaBuzz on Unsplash