Why your CRM never took hold — and how to make sales reps want to work in it

A CRM fails to take hold for one reason: it adds work for the sales rep and gives nothing back. A system like that gets sabotaged whatever the fines and stand-ups — and the numbers confirm it. According to Forrester, 38% of the problems in CRM projects come from human factors: slow adoption of the system, weak training. Technology itself accounts for 35%.

Below is what sabotage looks like in practice and why it is rational from the rep's side. And most importantly: how to redesign a CRM so that people start working in it without being forced.

Why don't sales reps keep the CRM up to date?

Sales reps don't keep the CRM up to date because it is built as a tool of control over them, not as a tool for their work. The system demands data entry, reporting and “maintaining records”, and in return offers only the sense of being watched. A rational person does not accept that deal: they do the minimum to be left alone, and sell the way that suits them.

Scott Edinger, a sales consultant, put it this way in Harvard Business Review:

“CRM systems are too often used for inspection — to report on progress of sales — rather than improving the actual process of selling” — Scott Edinger, Why CRM Projects Fail and How to Make Them More Successful, Harvard Business Review, December 2018.

In his surveys among clients, nearly 90% of executives answer “no” to the question of whether their CRM helps grow the business. The reason lies in the original framing: the system was bought to “see what sales is doing”, and nobody asked what it would give the people who work in it.

A Forrester survey breaks the problems of CRM projects down by source: 38% people (slow adoption, lack of training and change management), 35% technology, 33% processes. Resistance from people is the single largest loss item, larger than bugs and bad configuration. Fines do not close it: people start imitating use of the system, and the data in it becomes even more useless.

What does CRM sabotage look like in practice?

Sabotage almost never looks like open refusal. More often it is quiet workarounds: deals live in personal messengers, fields are filled in a minute before the stand-up, and a “shadow” spreadsheet grows alongside the CRM, where the rep does the real work. Formally the system is in use — in fact it is empty, and reports built on it are more dangerous than no reports at all.

The typical scenarios from our audits repeat from company to company:

• A client writes to the rep on their personal WhatsApp. For the company the deal does not exist until an invoice is needed.
• The “comment” field is filled with the words “ok”, “call”, “thinking it over”. The negotiation history lives in the rep's head.
• Deals are entered retroactively on the last day of the month — once it is already clear what closed. The pipeline turns into a chronicle.
• The best salesperson keeps “their own” spreadsheet and treats the clients as a personal book of business. When they leave, they take it with them — there is a separate breakdown of that scenario, when a business should move off Excel.

Russian data confirms the scale of the workarounds: according to a J'son & Partners survey (1,000 companies, winter 2025/26), 98.7% of companies use other tools in parallel with their CRM — chats, spreadsheets, notepads. The only question is what counts as the deal's main home: the system or the rep's personal correspondence.

How much time does the CRM take away from selling?

According to Salesforce State of Sales, reps spend only 28% of the working week on selling itself — the rest is eaten by data entry, approvals and reports. Manual entry is the main irritant here: 23% of CRM users name it as the main problem of working with the system. The rep feels this arithmetic every day, even without counting it in percentages.

Simple math. A rep has finished a conversation with a client — and spends 15 minutes “putting it into the system”: the record, the fields, a comment, a task. Ten contacts a day is two and a half hours of pure data entry. A system that adds that much work will be sabotaged, and in its own way that is fair.

That also sets the direction of the repair. As long as data entry is the price the rep pays for the manager's transparency, that price has to be driven down to zero, and the benefit of the system handed to the rep.

Three reasons a CRM dies within six months

The dead systems we have taken apart were killed by the same three things: the CRM duplicates work instead of removing it; the pipeline in the system does not match the real course of a deal; data gets in only by hand. Any one of these lowers the chances of adoption. All three together make failure a matter of time.

1. Duplicated work. The rep has sold — and then separately “enters the sale”. The CRM lives next to the work and demands that what has already happened be copied into it. Every such transfer is a tax on the deal.
2. A pipeline “like everyone else's”. The stages were invented during implementation from the integrator's template, while in real life deals move differently. The rep cannot reflect reality honestly in the system — and stops trying.
3. Manual entry. Everything entered by hand is entered with errors, late, or not at all. A report built on such a database looks like data, though the data left it long ago.

The scale of the problem has been measured: according to the CRM Failure Report 2025, 55% of CRM implementations do not meet their stated goals, and one project in ten is cancelled before launch. Only a quarter of projects hit goals, deadline and budget at the same time.

All three reasons are closed at the design stage, before the software is chosen. That is why in custom development we start CRM projects with an audit of the sales process: we look at how the deal actually moves, and only then decide what to build.

What does “a CRM built around the process” mean?

A CRM built around the process is designed in reverse order. First you record how the deal actually moves and what data is produced at each step. Then you build an interface in which that data records itself — with no separate “entering it into the system”. The rep gets a tool that saves them time from the first day, and the question of “how do we make them keep the CRM up to date” is removed.

The check question for every field and every screen: what does it give the person filling it in? If the answer is “nothing, it's for the report” — the field will be filled with garbage. The data for the manager is assembled by the system from events, while the rep is shown only their own work: deals, today's tasks, the client's history. Without 40 fields kept “just in case”.

How does data get into the CRM without the rep?

Four ways: inquiries arrive in the system on their own from every channel, statuses move on events in connected systems, documents are generated from deal data, and the routine of first contact is taken over by an AI agent. Manual entry stays where there is no doing without a human — in the substance of the negotiation. Everything else the machine records more accurately and faster.

• Inquiries. From the website, messengers, email and ad platforms, an inquiry lands in the CRM automatically — with its source and full history. The rep's personal WhatsApp stops being a sales channel.
• Statuses. An invoice is issued — the status changes by itself. A payment arrives from the bank — the deal is closed. That is the work of integrations: how the CRM, 1C (the accounting and ERP platform most Russian companies run on), the bank and the warehouse are linked together is covered in the breakdown on a single loop of systems.
• Documents. Contracts, invoices and acceptance certificates are generated from deal data in one click, with no manual filling of templates.
• First line. A first-line AI agent answers routine questions, qualifies the inquiry and creates the deal itself, with fields and the history of the correspondence. The rep joins a warm client, and the most tedious part of the entry — “what is needed, when, what budget” — is already done.

When the system saves the rep an hour a day, it does not need to be imposed. People start working in it because things are worse without it: inquiries are visible, documents take one click, the client's history is at hand.

What does the owner get from a living CRM?

Transparency without meetings. The pipeline is built on real data: how many leads there are, where they stall, what conversion each rep has. Sales plan and actuals update themselves. Every client's history sits in one place, and a rep leaving no longer means losing the client base. All of it is a side effect of a system that reps find convenient to work in.

Many know the cost from bitter experience: one company in five in the J'son & Partners survey implemented a CRM only after losing its deal data — that is, having paid for the lesson with its client base.

A separate screen for the decision-maker — sales, money, receivables in real time instead of a report for last month. Which numbers to put on such a screen and how not to drown in metrics is covered in the article on the owner's dashboard.

Off-the-shelf CRM or custom: which to choose?

By calculation, without ideology. Off-the-shelf systems work well on a standard sales process: lead, call, invoice, payment. A custom CRM is justified when the process is non-standard: a specific pipeline, your own entities — properties, batches, trips — deep integrations with accounting and production. The choice comes down to the question of which is more expensive: adapting the box to your process, or building a system around it from the start.

The Russian market is arranged predictably: in the same J'son & Partners survey, 49% of companies use Bitrix24, a Russian CRM and collaboration suite, 20% use 1C:CRM, and 9% amoCRM, another Russian CRM. For a standard process that is a workable choice.

We do both: we configure and extend off-the-shelf systems, and when the box starts getting in the business's way, we design a custom one. A fast start is covered by ready-made solutions from the catalog, each adapted to the process. The tool is chosen to fit the task.

How do you revive a CRM that is already a mess?

Tearing it down and implementing again is usually unnecessary — rebuilding what you have is cheaper. The working sequence: an audit of how the reps actually sell, a pipeline built around that process, a cleanup of excess fields and automated data entry. The system comes alive when entering a deal becomes faster and more useful than hiding it.

1. Find where the deals actually live: private chats, email, spreadsheets. That is the map of what the system is missing.
2. Rebuild the pipeline around the real course of a deal. The stages should be named the way the reps themselves name them.
3. Delete the fields nobody uses. Every required field has to answer the question of who needs it and what for.
4. Automate entry: inquiries from channels, statuses from events, documents from deal data.
5. Move the work over in stages and check the numbers: the pipeline report has to match the money in the bank.

This is a particular case of the rule that you cannot automate chaos: first the process, then the system. How that rule works in any part of a business, from inquiries to reporting, is covered in the basic breakdown business automation: where to start.

How do you know the CRM has taken hold?

By the behavior of the reps. A living system is visible after a week of observation: deals are entered on the day the inquiry arrives, the negotiation history reads without anyone retelling it, reps ask for improvements themselves instead of finding workarounds. And the main test: the pipeline report matches the money in the bank with no manual reconciliation. If the test fails, the system is still living apart from sales.

If your CRM never took hold and the deals have spread back across chats — tell us about your process. We will run an audit, find where the system adds work, and redesign it around how your reps actually sell: under contract, with the data kept in Russia and support after handover.

Sources

• Forrester: CRM Success Requires Focus On People, Not Only Technology — the sources of problems in CRM projects, 2016
• Salesforce, State of Sales (fifth edition): reps spend 28% of the week selling
• Scott Edinger, Why CRM Projects Fail and How to Make Them More Successful, Harvard Business Review, December 2018
• Johnny Grow, CRM Failure Report: 55% of implementations do not meet their goals, 2025
• J'son & Partners and Bitrix24: how companies choose a CRM, a survey of 1,000 companies, 2026
• CRM.org: 45 CRM Statistics — manual data entry and adoption barriers