Write your sales process down first, then let the CRM enforce the parts that matter. Borrow the tool's defaults only where you have no strong opinion of your own.
- Process first, configuration second
- Adopt defaults where you have no view, adapt where you win deals
- Every required field must earn its place
Every small team that buys a CRM hits the same question within a week. The tool comes with stages, fields and a pipeline view that look sensible. Your team already has a way of selling that mostly works. Which one gives way?
The honest answer is: both, in a specific order. Your process decides what needs to happen. The CRM decides how consistently it happens. Get that order wrong and you end up either fighting the software every day or selling the way a product manager in another country imagined you sell.
The case for letting the CRM lead
Default pipelines exist for a reason. They encode a rough consensus about how B2B and service sales move: a lead appears, someone qualifies it, a conversation turns into an opportunity, a proposal goes out, the deal closes or dies. If your current process lives in one person's head and a shared spreadsheet, the default is probably better than what you have.
Letting the tool lead also has a speed advantage. You can be live in an afternoon. Reports work out of the box. New hires see a structure that matches what they used at their last job. For a team of two or three people with no formal process, that is a reasonable trade.
The risk is that defaults are generic by design. They cannot know that your deals stall at site visits, or that your best customers always come through referrals and skip qualification entirely. A generic pipeline will happily record your sales. It will not help you win more of them.
The case for letting the process lead
If you already know what separates won deals from lost ones, that knowledge belongs in the CRM. A stage called Demo booked means nothing if your real turning point is Decision maker on a call. The pipeline should mirror the moments where deals actually change state, not the moments that are easy to log.
Process-led setups also produce better data. When stages reflect real buyer commitments, conversion rates between stages tell you something. When stages reflect internal activity, conversion rates mostly tell you how diligent people are with data entry.
The risk here is over-engineering. Teams that start from their process often turn every habit into a mandatory field. Six months later, nobody fills them in, and the CRM is a museum of good intentions.
A practical order of operations
Here is the sequence that works for most small teams we have seen.
- Write the process on one page. List the steps a deal goes through, in plain words, before you open the CRM settings. If you cannot fit it on a page, it is too complicated to enforce.
- Mark the commitments. Circle the steps where the buyer does something, not you: agrees to a call, shares a budget, introduces a colleague, signs. These become your stages.
- Compare with the defaults. Where the CRM's stages match your commitments, keep them and their names. Where they do not, rename or replace them.
- Choose required fields last. For each stage, ask what you need to know to move to the next one. Make only that required.
- Run it for 30 days, then prune. Delete any field that is empty on most records. Merge any stage that deals pass through in under a day.
Adopt, adapt or ignore
Not every part of the CRM deserves the same treatment. A simple way to sort decisions:
| Area | Usually adopt | Usually adapt | Usually ignore at first |
|---|---|---|---|
| Pipeline stages | Lead, Won, Lost | Middle stages to match buyer commitments | Sub-stages and probability tweaks |
| Contact fields | Name, email, phone, company | Source, role in the decision | Social profiles, birthdays |
| Activities | Calls, emails, meetings | Task types for your follow-up rhythm | Detailed time tracking |
| Reports | Pipeline value, win rate | Stage-to-stage conversion | Custom forecasting models |
| Automation | Reminders on stale deals | Handoff from marketing to sales | Complex multi-branch workflows |
The pattern is consistent. The edges of the pipeline (how a lead arrives, how a deal ends) are universal, so the defaults are fine. The middle is where your business is different, so that is where your own process should win.
Where software should push back
There is one area where the CRM should dictate, and you should let it: discipline. A good tool makes it hard to leave a deal without a next step. It flags opportunities that have not moved in two weeks. It asks for a loss reason when you close something as lost.
These are not process decisions. They are hygiene. Most teams already agree with them in principle and skip them in practice. Letting the software enforce them is the cheapest coaching you will ever buy.
This is also where a command bar helps. In Salesformics, the GX Box lets you type next step call Friday or close lost price instead of opening a form. When logging the right thing takes three seconds, people stop arguing with the process.
Warning signs you have the balance wrong
- Reps keep a private spreadsheet. The CRM does not reflect how they actually work, so they built their own.
- Every deal sits in the same stage. Stages are too broad, or nobody knows when to move a deal.
- Forecasts are always wrong in the same direction. A stage is counted as a commitment when it is really hope.
- Managers ask for updates in meetings. If the CRM held the answer, nobody would need to ask.
- Fields are filled with placeholder values. A required field nobody needs gets typed in as "n/a" or "tbc".
Each of these has the same fix: go back to the one-page process, check which step the CRM is misrepresenting, and change the configuration rather than the people.
A worked example
Take a small agency selling website projects. The default pipeline has Qualified, Demo, Proposal, Negotiation. The agency never does demos. Its real turning points are a discovery call, a scoping workshop the client pays for, and a fixed quote.
So the pipeline becomes Enquiry, Discovery call, Paid workshop, Quote sent, Won, Lost. Required fields: budget range at discovery, workshop date at workshop, quote value at quote. Nothing else is mandatory.
Within a month, the agency can see something it only suspected before: clients who pay for a workshop convert to projects far more often than those who go straight to a quote. That single insight changes how they sell. The default pipeline would never have shown it, because it had no stage for the workshop.
The short version
Do not let a CRM invent your sales process, and do not ask it to reproduce every habit you have. Write down the moments where buyers commit, build stages around them, adopt the defaults everywhere else, and let the software enforce the boring discipline that keeps data honest. Review it monthly. The process will change, and the CRM should change with it.
Questions
How many pipeline stages should a small business have?
Usually five to seven, including Won and Lost. Fewer and you cannot see where deals stall. More and people stop moving deals between stages accurately.
Should different products have different pipelines?
Only if the buying journey is genuinely different, for example a one-off sale versus a subscription with a trial. If the steps are the same, use one pipeline and a product field.
How often should we change the CRM setup?
Review monthly for the first quarter, then quarterly. Change stages rarely, because it breaks historical reports. Change fields and automations as often as you need.
