Do CRM projects fail? Where the 50% figure comes from

The claim that “30% to 70% of CRM projects fail” has circulated for more than twenty years, and almost none of its versions leads to an original document. The most cited verifiable figure comes from a Gartner press release from June 2003, which said roughly 50% of CRM projects failed to meet executive management expectations, and that definition of failure is much narrower than the sentence that travels through slide decks. The 55%, 63% and 70% that show up in presentations arrive second or third hand, with no sample and no method. Peer-reviewed academic research with real samples tells a more useful story for leadership: CRM has a positive average effect on performance, and that effect depends on support from the executive team, on use by salespeople and on the information routines the company builds around the system.

I hear the failure figure in almost every diagnostic meeting, either as an argument against implementing or as an excuse for the project that went wrong. After more than 200 companies served, I distrust it for the same reason I distrust any statistic without an origin: nobody in the room knows what was measured. I went after the sources to separate what holds up from what became folklore, and I wrote this for the executive team about to approve a CRM project, for the leaders who answer for its use and for the HR team that organizes training.

Where do CRM failure statistics come from?

Most of the CRM failure statistics in circulation were born between 2001 and 2003, at the end of the first wave of large implementations. Context matters, because CRM at that time required in-house servers, heavy customization and custom-built integration, and the Bain authors themselves estimated at least 24 months to implement. Even so, the figures are still applied to the Brazilian company that signs up for a cloud platform in 2026.

What can be verified in the original

The first figure with an identifiable origin is Bain & Company’s, published in a Harvard Business Review article by three of the firm’s consultants (Rigby, Reichheld and Schefter, 2002). In Bain’s 2001 management tools survey of 451 senior executives, CRM ranked in the bottom three of 25 tools for satisfaction, and one in five users reported that their CRM initiative had not only failed to deliver profitable growth but had also damaged long-standing customer relationships. The figure measures executive perception, and the sample is described only by the number of respondents, but the source and the method are stated.

The second is Gartner itself. In a June 2003 press release, reproduced in full by Hospitality Net, the firm said that “roughly 50 percent of all CRM projects fail to meet the executive management expectations” and predicted that, through 2005, 75% of CRM projects that did not deliver measurable ROI would have failed because of poor business executive decision making (Gartner, 2003). The release promoted a paid 354-page report, closed to the public, and the open text gives neither sample nor method. Note also that the 75% is a prediction about a subset, the projects without measurable ROI, and does not work as a market failure rate.

What circulates without an original document

The same HBR article opens with another figure: 55% of all CRM projects don’t produce results, “according to Gartner Group.” The authors do not say which document, and I could not find the original. The article also cites a CRM Forum poll, with no sample given, in which 87% of managers blamed failure on a lack of adequate change management and 4% on software problems. The “Gartner 2001” 50%, the 63% attributed to a 2013 Merkle Group study and the 70% attributed sometimes to Gartner, sometimes to other consultancies appear in blogs, trade press articles and vendor pages that cite one another. I looked for the original report behind each one and found none published with sample and method.

Figure in circulation	Usual attribution	What I found	Status
1 in 5 users whose CRM damaged customer relationships	Bain, 2001	Survey of 451 executives, cited by the Bain authors in HBR	Verifiable in the authors’ publication
~50% fail to meet executive expectations	Gartner, 2003	June 2003 press release, no sample disclosed	Verifiable in the release
75% of projects without ROI will fail by 2005	Gartner, 2003	Prediction in the same release	Verifiable, but a prediction
55% don’t produce results	Gartner, cited in 2002	Citation in HBR with no document	Second hand
87% due to lack of change management	CRM Forum	Poll cited in HBR, no sample	Second hand
50% (2001), 63% (2013), 70%	Gartner, Merkle Group and others	Blogs and articles with no link to a report	No primary source found

How does the definition of failure change the number?

CRM failure can mean very different things, and each definition produces its own rate, as the table below shows. Gartner’s 50% measures executive expectations, which depend as much on what the executive team expected as on what the project delivered, and an expectation inflated during the sale is enough to turn a reasonable project into a failure statistic.

The study by Becker, Greve and Albers shows this effect with data. The authors sent a questionnaire to 400 companies with CRM experience, selected from the client base of an international consultancy, in ten European countries and four industries, and received 90 usable responses, in more than 69% of cases from senior executives in top management or in marketing and sales (Becker, Greve and Albers, 2009). The conclusion, in the version the authors themselves published for managers, is that the combination of poorly implemented projects and overly high expectations produces what companies perceive as failure (Becker, Greve and Albers, 2010). Neither form of implementation, technological or organizational, had a significant effect on customer retention. An executive team that bought CRM to reduce churn and judges the project by that indicator will conclude it failed, even if acquisition improved.

The chosen metric also decides the verdict. In a study of a large sample of US commercial banks, CRM implementation was associated with a decline in cost efficiency and an increase in profit efficiency (Krasnikov, Jayachandran and Kumar, 2009). The same bank looks like a failure to whoever reads the expense line and like a success to whoever reads the margin. The authors also found that time since adoption improves performance, and judging the project at month six by a third-year yardstick distorts the verdict.

Definition of failure	Question it answers	Common misreading
Schedule and budget	Did the project go as planned?	A late project that later paid off counts as failure
Abandonment	Did the company switch the system off?	A system left on and unused counts as success
Adoption	Does the team work in the CRM?	Many logins with empty records count as adoption
Executive expectation	Was the executive team satisfied?	An unrealistic expectation becomes a delivery failure
Business result	Did the chosen metric improve?	Early cost hides the margin gain

Does CRM improve results when it is implemented well?

Academic research tends to answer yes, with a moderate and conditional effect. In the Becker study, technological implementation had a moderate impact on customer acquisition and on customer maintenance, which includes satisfaction and cross-selling. Organizational implementation, with segment-based structure, incentives and training, had no significant effect on its own, and gained one when top management support was present. Technology, in turn, had a much stronger impact when employees accepted and used the system. In the bank study, profit efficiency went up. The link between CRM and commercial results gets its own reading in CRM and performance.

The reading I take to the executive team is that the tool delivers a small part of the result and the people around it deliver the rest. In my experience, that explains why two companies in the same industry, on the same platform, reach opposite results: CRM amplifies the management routine it finds, including a disorganized one.

Why do CRM projects fail, according to research?

The causes research supports are few, and they show up in studies from different periods and countries. I separate each one by the evidence behind it, because some rest on sampled studies and others on consultant opinion.

Rejection by salespeople after training

Speier and Venkatesh followed 454 salespeople at two firms that had implemented sales force automation tools, systems used to support CRM strategy in the sales function (Speier and Venkatesh, 2002). Right after training, salespeople had positive perceptions of the technology. Six months after implementation, the technology had been widely rejected, absenteeism and voluntary turnover had increased, and job satisfaction and organizational commitment had dropped at both firms. Salespeople with stronger professional commitment were the ones whose perceptions worsened most as they used the tool.

This result undermines the training evaluation done on the way out of the room, which measures the good impression of the day while rejection only shows up in month six. The behavioral causes and the 30-day plan to change usage are in CRM adoption.

Lack of top management support

In the Becker study, organizational change affected performance only when top management visibly supported CRM, with involvement in the project and a communicated strategic orientation. Gartner’s 2003 release points the same way, but as an analyst prediction, with no published data. The practical implication falls on the sponsor: the executive who backs the project and keeps asking for the report in a spreadsheet cancels out the investment in training and incentives.

Tool before strategy and process

The Bain authors in HBR, who analyzed customer loyalty initiatives at more than 200 companies over ten years, argue that many executives treat CRM as software that manages customer relationships for them, and they list as pitfalls implementing CRM before creating a customer strategy, installing the technology before reorganizing processes, metrics and compensation, and assuming that more technology means more results. This is consulting evidence, based on cases, without a statistical sample, and it is consistent with the study by Jayachandran and colleagues, which, using a diverse sample of firms, found that organizational routines for processing relational customer information explain customer relationship performance and that CRM technology moderates that effect (Jayachandran et al., 2005).

Data and information routines

The Jayachandran study measures relational information routines, a broader concept than the quality of the customer database. Becker and colleagues argue that CRM effectiveness depends on the quantity and quality of the data entered, but they treat this as a premise of the model, without measuring the effect. The specific evidence on data quality as a cause of CRM failure is weaker than the evidence on adoption and leadership. How to measure the database and what pollutes it are covered in CRM data quality.

Expectations misaligned with the objective

CRM serves acquisition, maintenance and retention differently, as the European and bank studies showed. An executive team that sets the project’s objective before buying, with an indicator and a deadline, lowers the chance of calling a project a failure when it delivered what it could.

Cause	Main evidence	Type of evidence	Strength
Rejection and low use by the team	Speier and Venkatesh, 2002: rejection six months after implementation; Becker, Greve and Albers, 2009: technology works better with employee acceptance	Study at two firms and survey of 90 companies	High for sales forces
Lack of top management support	Becker, Greve and Albers, 2009: organizational change only works with management support	Survey of 90 European companies	Moderate
Tool before process and strategy	Rigby, Reichheld and Schefter, 2002; Jayachandran et al., 2005	Consulting cases and academic survey	Moderate, with converging sources
Data and information routines	Jayachandran et al., 2005	Academic survey; information routines, broader than data quality	Moderate for routines, weak for data quality alone
Misaligned expectations	Becker, Greve and Albers, 2009; Krasnikov, Jayachandran and Kumar, 2009	Survey and US bank data	Moderate
Change management	CRM Forum poll cited in HBR	Poll with no sample disclosed	Weak

Summary of the research cited

Source	Year	Sample and method	Main finding
Rigby, Reichheld and Schefter (Bain), HBR	2002	Bain 2001 survey of 451 executives; analysis of more than 200 companies	CRM in the bottom 3 of 25 tools for satisfaction; 1 in 5 users reported harm to customer relationships
Gartner, press release	2003	Analyst research, no sample disclosed	~50% of projects fail to meet executive expectations
Speier and Venkatesh, Journal of Marketing	2002	Survey of 454 salespeople at two firms, before and six months after	Positive perceptions after training and wide rejection six months later
Jayachandran et al., Journal of Marketing	2005	Survey of a diverse sample of firms	Information routines explain performance; technology moderates the effect
Becker, Greve and Albers, IJRM	2009	90 companies in ten European countries and four industries	Implementation effect depends on management and employee support; no effect on retention
Krasnikov, Jayachandran and Kumar, Journal of Marketing	2009	Data from a large sample of US commercial banks	CRM lowers cost efficiency and raises profit efficiency; the effect improves with time in use

What these data do not say

The studies with sample and method are old. The main ones came out between 2002 and 2009, when CRM ran on the company’s own servers, and I did not find a recent academic study with the same design showing whether cloud platforms changed the rates. None of them was done in Brazil, and the average Brazilian company differs in size, process maturity and sales culture from the European firms with more than 5,000 employees that dominate the Becker sample.

Performance measures are largely self-reported. In the European study, the people who rated the improvement were executives responsible for the CRM itself, and self-assessment of one’s own project tends to be generous. The bank study uses objective financial data, but covers a single industry. The surveys show an association between management support, use and performance, and association does not prove causation: better-run companies may have more executive support and better performance for the same reasons.

The Speier and Venkatesh study followed two firms and the tools of twenty years ago. I use it as a warning about the timing pattern, with a good impression at training and a drop months later, without treating it as a rate for any company. Finally, none of the sources allows anyone to say what percentage of CRM projects fail today. Anyone stating a precise number is repeating a citation.

What leadership does with these numbers

Use the failure figures as a list of risks the executive team can control before signing the contract.

1. Write the definition of success before buying. Decide which objective the project serves first, acquisition, maintenance or retention, with an indicator, a baseline and a deadline.
2. Name an executive sponsor with time on the calendar. The support research measures is visible: presence in project decisions, communication to the team and use of the CRM in the meetings the executive team itself runs.
3. Design process and handoffs before configuring. Agree on the definitions of lead, opportunity and at-risk customer and on what each handoff between teams must contain, as in the SLA between marketing, sales and customer success.
4. Measure usage through month six. Track records per person, deals with a next step and won deals created before the close date for at least six months.
5. Bring management into the system. Run the pipeline review and the account review with the CRM on screen, and retire the parallel spreadsheet on a set date.
6. Judge results in the right window. Separate cost and margin in the analysis and wait at least one full sales cycle before concluding failure or success.

CRM, failure and L&D: what HR needs to assess

The Speier and Venkatesh finding matters a great deal to HR, because it shows that the reaction survey right after training can point the opposite way from what happens months later. The measure that counts is system use on the job, and it has to be collected in the CRM, by person and by role, before training and at one, three and six months. It is worth cross-checking absenteeism and turnover against the project schedule, because both rose after implementation in the same study. The roles that record data in the CRM take part in the training and, before them, the leaders who run the meetings, because the effect of organizational training depends on management support.

If your company is about to implement a CRM, or already has and is watching the team drift back to spreadsheets, that is the work of the in-company CRM adoption training. I run the program with the leaders and with each role, in the CRM you already use, and HR receives the usage indicators to compare before and after. The group leaves with the CRM best practices guide, with the shared definitions and what each role records, and with the account and database management manual, with the routine that keeps the system in use after go-live.

Frequently asked questions

What is the failure rate of CRM projects?

I did not find a current rate measured with a public method. The most cited verifiable figure comes from a 2003 Gartner press release, which said roughly 50% of projects failed to meet executive management expectations. The 55%, 63% and 70% versions circulate without an original report that can be located.

Where does the 70% CRM failure figure come from?

The 70% appears attributed to Gartner and to other consultancies in blogs and vendor pages, with no link to the original document. I did not find any published report with sample and method that supports it.

Why does CRM not work in so many companies?

Academic research points to rejection by salespeople months after implementation, lack of visible top management support, deploying the tool before the process and the strategy, and expectations misaligned with what CRM delivers. The strongest evidence concerns employee use and management support.

Does CRM improve company results?

Studies with samples find a moderate positive effect, conditional on management support and team usage. In the US bank study, cost rose at first and profit efficiency increased, with larger gains the longer the system was in use.

Does switching CRMs fix a failed project?

In most cases, no, because the causes research supports lie in people, processes and leadership, and they migrate along with the company. Switching makes sense when the operation already uses the system with discipline and the tool limits the sales model.

How do you prepare the team so the failure does not repeat?

Train each role in the company’s own CRM, with real records, and prepare the leaders who run the meetings first. Measure system use before training and for at least six months after, because rejection tends to appear when the project has already been declared delivered.