CRM for startups: what to track before you have a sales process

A CRM for startups is usually bought too early in the wrong shape: an enterprise pipeline with seven stages, forecast fields and lead scoring, adopted by three founders who have not yet closed ten deals. Within a month the stages are fiction and the data is stale, because the process they describe does not exist yet.

Early-stage sales is a learning exercise. The record that matters is not "which stage is this deal in" but "what did we learn from this conversation, and what did we promise".

What a startup actually needs to record

• Who you spoke to, at what company, and how they found you. Source is the field founders skip and later desperately want.
• What problem they described — in their words. This is your product research, and it is worth more than any stage label.
• What you promised, and by when. Broken promises at ten customers become a reputation at a hundred.
• Objections, verbatim. The repeated ones are your pricing page, your FAQ and your roadmap.
• Outcome and reason. Not just won or lost — why, in one sentence a stranger could act on.
• Next action with a date and an owner. Everything else is optional; this one is not.

Before you have a process, the pipeline stages are the wrong abstraction. Two lists carry you a long way: conversations happening now, and conversations that need a next step. Add stages when you can name them from what actually happened, not from a template.

When to add structure

The signals that you are ready for a real pipeline are concrete: the same sequence of steps has repeated across at least a handful of deals, more than one person is selling, or you need to forecast because someone is planning hiring against it. Until then, structure is overhead that hides the learning.

The same applies to fields. Every required field taxes every interaction. Start with three — owner, next action, next date — and add a field only when you can name the decision it will inform.

Setting it up in an hour

1. Import whatever exists: the spreadsheet, the inbox contacts, the conference list. Deduplicate before anyone starts working, not after.
2. Add one custom field for source and one for the problem stated. These two carry most of the early learning.
3. Log conversations the same day, in three lines: what they said, what you promised, what happens next.
4. Tag by segment guess rather than by industry taxonomy — you are testing hypotheses, and the tags should be the hypotheses.
5. Review weekly: anything with no next action, anything promised and not delivered, anything that went quiet.
6. Every month, read the last twenty problem statements in one sitting. Patterns show up in bulk that are invisible one call at a time.

Ettex CRM fits that stage: contact profiles with phones, emails, addresses and custom fields for source and problem, groups and tags for segment hypotheses, notes and activity logging so calls and meetings sit on one screen, a visual pipeline for when you do have repeatable stages, change history showing who edited what and when, shared address books so co-founders see the same record, duplicate merging after a messy import, inline editing straight in the list, instant search by name, email, phone or note, and import and export as vCard or CSV. Deals link to the documents and invoices behind them, which matters the first time a customer asks what exactly they signed.

Startup CRM mistakes

• Copying an enterprise pipeline template and pretending the stages are real.
• Tracking activity counts. At ten customers, calls made is not a metric, it is a distraction.
• Keeping customer knowledge in one founder's inbox, which becomes unreachable the moment they are on a plane.
• Deleting lost deals. The reasons they were lost are the most valuable dataset an early company has.
• Choosing a tool whose export is worse than its import — the switch later costs the history.

A CRM for startups should record what you are learning, not enforce a process you have not discovered yet. Owner, next action, next date — and the customer's own words.