A CRM does not die. It empties out

A CRM does not die on launch day. It empties out, week after week, and nobody ever decides to abandon it. That is what makes the failure invisible until the quarter when the CEO stops believing the report. Part 1 laid the foundations, worth reading first on kasar.app/en/guides/crm-pme-fondations: a defined process, data entry removed, a minimal data model. What happens now is a different kind of problem, because it is no longer about configuration.

The short answer: adoption is not decreed, it is measured, and what people usually measure proves nothing. A login rate does not tell you whether a rep is working inside the tool. Two indicators are enough: the percentage of reps logging at least three activities a week, and the percentage of open deals carrying a next action and a close date that has not passed. The target we recommend is roughly 85% at thirty days. The rest of this guide covers the levers that get you there, the seven mistakes that take you away from it, and the schedule that fits in a month.

01A CRM does not die, it empties out

The mechanism is always the same, and it is gradual. Week 1, everyone logs. Week 3, the busy days get skipped. Week 6, two reps have reopened their old file, "just for themselves". Week 10, the pipeline review is run from memory, because the screen no longer reflects reality. At that point the weekly report is wrong, the CEO knows it, and stops asking for it. Nobody decided anything, and the project is dead.

What triggers the decay is never a missing feature, it is an imbalance: the tool asks more than it gives back, to the person doing the typing. Every lever below corrects the same balance: the rep pays for the data entry, the CEO collects the visibility. As long as that balance stays negative for the person typing, no rule and no dashboard will hold past the week 6 described above. So they start from them: make them win first, ask them for something second.

02Start with a pilot, not a memo

A CRM designed with the sales team gets adopted by the sales team; a CRM decreed from above gets boycotted in silence. The resistance is not obstruction: the veteran who keeps everything in their head and in their own file is protecting their autonomy and control over their accounts. You do not win that fight through authority, you win it by making the tool useful to them in the first week. The most effective shortcut is a pilot, on one or two reps, before any wide rollout.

03Sponsorship, ownership and the weekly review

The pilot creates appetite, it does not create legitimacy. That comes from the top, and it cannot be delegated. A project delegated to "someone in operations" dies for lack of authority; a project the CEO visibly uses becomes legitimate. But everyone still needs to know what they own.

• The rep owns their records : accountable for their contacts and their deals, not for reporting to the boss.
• The manager owns the decisions : deal progression, trade-offs, pipeline hygiene before the review.
• The CEO owns the process : the stages, the required fields, and the definition of what adoption means.
• A named point person : single contact for every question during weeks 1 to 6.

Sponsorship shows up in exactly one ritual, and it is non-negotiable. Every week, the CEO or sales director runs a short pipeline review, 15 to 30 minutes, screen shared, directly in the CRM. Only open deals are covered, from the most advanced to the most uncertain. Any deal with no activity for 7 days is flagged stalled and handled first. Seven days and not fourteen: a managed team reviews its pipeline every week, a founder working alone does not. Everyone leaves with a next action and a date. And before walking the deals one by one, the manager announces a single number: total open pipeline divided by what the team has to sign this quarter. The threshold is not universal, it follows from your win rate: at 33% conversion you need 3x the target, at 20% you need 5x. Below your own threshold, the review will not help until the top of the pipeline is refilled.

04Zero double entry: capture the real channels

The weekly review is only worth anything if the screen tells the truth, and the screen only tells the truth if the channels feed into it. SMB selling happens on the phone, by email, and increasingly on WhatsApp and LinkedIn. If those channels live outside the CRM, the tool stays a partial, stale copy, and every update becomes double entry again. That is exactly the imbalance that empties a CRM.

The choice of tool matters less than people think. A salesperson never abandons a CRM because a feature is missing. They abandon it because the tool asks them to retype what they just did in their inbox. The one differentiator that genuinely moves adoption is therefore elsewhere: how much manual entry the tool removes.

05Measure adoption, not logins

The levers above produce an effect, and you still have to observe it. This is where most rollouts pick the wrong instrument: they track logins. A login proves no usage at all. It proves a tab stayed open.

These two indicators have a useful property: they degrade before your sales results do. The day a growing share of your open deals no longer carries a next action, the quarter is already under way and the revenue will not say so for weeks. That is what makes them steerable, and it is what no login rate will ever give you.

06How long does it take to roll out a CRM in an SMB?

07The 7 mistakes that kill adoption

1. 1Tool before process : choosing or configuring the software before defining the process means digitising the chaos.
2. 2Over-configuring on day 1 : 40 fields, custom objects, required fields outside the real workflow, so friction then abandonment.
3. 3Forcing instead of removing : you do not manage a rep into paying a data-entry tax, you automate the capture.
4. 4No sponsorship : a project delegated to operations, with no authority, dies.
5. 5Migrating dirty data : dumping an old spreadsheet into the new tool kills trust on day one.
6. 6No integrations : a CRM disconnected from email and phone means permanent double entry and stale data.
7. 7Training once then disappearing : with no point person, no follow-up in weeks 1 to 6 and no adoption measure, momentum fades.

08The data hygiene routine

09Your rollout and adoption checklist

One closing note, and it matters for the next twelve months. Everything above exists to get you a filled database, and a filled database is no longer just a condition for steering the business. It has become the entry condition for everything AI can do with your pipeline: prioritise, prepare a follow-up, flag what is slipping. An agent only knows what your CRM contains. The adoption you achieve in these thirty days decides what you will be able to automate next year. Before any of that, one number to take on Monday morning: count your open deals, then count the ones carrying a dated next action that has not passed. The ratio of the two is your starting point, and you can take it again in thirty days without us.