Why CRM projects fail
In the ones I have been called in to rescue, the software was working correctly almost every time. Fields saved. Workflows fired. Reports ran. And nobody was using it.
That is the thing to understand before you spend money. A CRM project does not fail the way a building falls down. It fails quietly, over about two quarters, as people drift back to the spreadsheet they trust. By the time anyone says the word failure, the decisions that caused it were made months earlier, most of them in the first fortnight.
Four patterns account for nearly all of it, and none of them are about which platform you bought.
Nobody owns it
This is the big one, and it is boring, which is why it gets skipped.
In week three somebody will ask a small question. Where do I log a call that came through my mobile. Do I create a new deal or reopen the old one. Why can I not see the Mumbai pipeline. These are not support tickets. They are thirty-second answers that need a person who knows both your business and the system.
If that person does not exist, the question does not get asked twice. The rep invents a workaround, the workaround spreads, and six months later your data is a record of four different conventions. The reports stop matching reality, managers stop trusting them, and once managers stop looking at the CRM, reps stop maintaining it. That is the whole death spiral and it starts with an unanswered question in week three.
The owner does not need to be technical. They need to know how the business works and have enough standing to settle an argument about how the system is used. Product knowledge takes a few weeks to acquire. Standing does not.
It was built for how management wishes people sold
Every discovery conversation has a moment where the process on the whiteboard and the process in the building turn out to be different things.
The whiteboard version is cleaner. It has the stages in a sensible order, the qualification step everyone agreed to in a meeting in 2023, the handover that is supposed to happen between pre-sales and delivery. The version in the building skips two of those, because skipping them is how deals actually close.
If the CRM gets built from the whiteboard, you have made the system an obstacle to doing the job. Reps will not stop closing deals the way that works. They will stop updating the CRM, which from their point of view is the lower-cost option, and they are right.
This is why I ask people to walk me through their last three deals rather than describe their process. Three real deals, including the one that went strangely, will tell you where the process is genuinely flexible and where it only looks flexible.
Sixty fields on the deal layout
Someone circulates a request for what people would like to report on, the responses get added up, and the Deal layout arrives with sixty fields on it, a third of them required.
Every required field is a tax on the person filling it in. And people route around taxes. They do not refuse, which would at least be visible. They type something: N/A, TBC, the first option in the dropdown, whatever clears the validation fastest.
Nine months later you are running board reports off that. The numbers look authoritative because they are in a system rather than a spreadsheet, and they mean very little. A field nobody fills in honestly is worse than no field at all, because an empty column is obviously empty and a column full of "TBC" looks like data.
Twelve fields people complete properly beat sixty that are half empty and quietly wrong. For every field on the list, ask who fills this in and what happens if they leave it blank. If the answer to the second is "nothing", it should not be required. If the answer to the first is "nobody", it should not exist.
Automation built before anyone used the system
Automation is the most satisfying part of an implementation to specify and the easiest to get wrong, because at kickoff you are automating a process that has never actually run in this system.
What you produce is a guess, hard-coded, and then defended, because somebody paid for it. When reality turns out differently, the automation becomes something people work around rather than something that helps.
Wait four weeks. Watch what people genuinely repeat by hand. Then automate that. The list you write after a month of real use is shorter than the one you write at kickoff and almost entirely different, and this is the single cheapest improvement available on most projects.
What the surviving projects have in common
A named internal owner with time actually allocated, not hoped for. A process documented from real deals rather than from the org chart. A deliberately short field list. Automation added in month two rather than week one. And training by role, so a rep learns how to work their day rather than taking a tour of the features.
None of that is expensive. All of it is easier to put in place before kickoff than after month three, which is the only real argument this page is making.
Questions I get asked
What share of CRM projects fail?
Published figures run from about 30 to 70 percent, which is a range wide enough to tell you the number is not worth much. Everyone defines failure differently. The useful question is narrower: is your team still using it in a year, and do your managers believe the reports.
Can a failed implementation be recovered?
Usually, and for less than people fear, because the configuration is rarely the problem. Recovery normally means cutting the field list back hard, re-agreeing the process with the people who actually sell, and appointing the owner who was missing. The rebuild is small. The conversations are the work.
Is switching platforms ever the answer?
Occasionally, when you are genuinely hitting a product limit. Far more often switching just moves the same four problems onto a new system, at the cost of a second implementation. Before you migrate, check whether anything on this page describes you, because if it does, the new platform will fail the same way.
Disclosure: This site is written and operated by Flowpeak Consulting, which sells Zoho implementation and rescue work. Weigh the advice accordingly.
Keep reading
What it costs
Hours rather than seats, and why cutting training is the most expensive saving available.
Read the guide →Choosing a partner
Who actually does the work, what the scope excludes, and what happens after go-live.
Read the guide →What the work involves
The phases in order, and why configuration is the small part.
Read the guide →