Sales Transformation
CRM transformation: a sales director's guide to doing it in the right order
10 mins read
KEY TAKEAWAYS
Most CRM projects disappoint because they start with the platform. The order that works is process, then data, then configuration, then AI.
The sales director owns the process: stage definitions, exit criteria, evidence standards and the management rhythm. IT and partners own the build.
Fewer fields, enforced, beat many fields ignored. Every required field should support a decision someone actually makes.
Adoption follows management. When pipeline reviews and forecast calls run from the CRM, the team keeps it current.
Judge the work by commercial results such as stage conversion, cycle length, win rate and forecast accuracy, not by logins.
Most CRM transformations fail for the same reason: they start with the platform. The forecast call where nobody trusts the numbers, the pipeline the team updates on Friday afternoon and the board asking whether a new system will fix it all point to the same gap. The CRM is being asked to enforce a sales process that was never designed.
The order that works is simple. Design the process first, then the data standards, then configure the platform, then add AI. Each step below covers what it involves and what the sales director should own, so you can run a CRM transformation that changes how the team sells instead of how it types. It's the approach behind our sales process and CRM transformation work.
What is CRM transformation?
CRM transformation is the work of changing how your organisation sells, then making the CRM enforce that way of selling. A CRM implementation configures a tool. A CRM transformation defines the sales process, the stages, the evidence each stage needs and the data leadership relies on, and only then builds the platform around them.
The difference matters because the two projects fail in different ways. An implementation can go live on time and on budget and still leave you with a pipeline nobody believes. A transformation is judged on whether forecasts become evidence-based and deals move faster.
Why so many CRM projects disappoint
You will often read that 70% of CRM projects fail. That figure has no traceable source. VentureBeat looked into it in 2015 and found that the report it is usually credited to never contained it. The older and more useful number comes from Gartner, which found in 2001 that 55% of CRM projects failed to meet expectations. As Gartner's analyst later explained, the press dropped the words "meet expectations".
The wording matters. Most CRM projects don't collapse; they disappoint, because nobody agreed what they were meant to achieve. Harvard Business Review named the root cause back in 2002: the assumption that CRM is software that manages customer relationships for you.
In our experience, the same four causes come up again and again:
No agreed sales process. The CRM has stages, but nobody has defined what has to be true for a deal to move from one to the next.
No definition of success. The project measures go-live and adoption, and nobody tracks forecast accuracy, conversion or cycle length.
Data nobody owns. Fields multiply, duplicates build up and every team keeps its own spreadsheet.
A system built for inspection. Sellers feel the CRM is there to check on them, so they update it as late and as little as possible.
None of these is a technology problem, and all of them land on the sales director's desk.
What the sales director should own, and what to hand off
The sales director should own the sales process and the decisions it supports. IT, revenue operations or an implementation partner should own the build. When this split is blurred, the platform ends up designing the process by default.
The sales director owns:
The stages of the sales process and the purpose of each one.
Entry and exit criteria, and the evidence a deal needs before it moves.
Qualification standards, including the method the team uses.
The forecast categories and what each one means.
The management rhythm: pipeline reviews, deal reviews and the forecast call.
IT, revenue operations or the partner owns:
Configuration, integration and data migration.
Automation that removes manual data entry.
Permissions, reporting builds and system administration.
Two decisions only the sales director can make: which information a deal must have before it is forecast, and what happens in a review when that information is missing. Every other design choice follows from those two.
Step 1: Decide what the CRM must help you decide
Start with the decisions leadership needs to make. The list of fields comes later. Which deals go in the forecast? Where should the team spend its time this quarter? Which accounts need senior attention? What is the pipeline really worth?
Forecasting is usually the sharpest test. When Gartner surveyed sales operations teams in 2020, only 45% of sales leaders and sellers had high confidence in their forecast's accuracy, and Gartner named poor data quality as one of the main causes. A CRM that can't support the forecast call is not doing its main job.
Write down the five or six decisions the CRM must support. For each one, note the information it needs and who provides it. This short list becomes the brief for everything that follows, and it stops the project from turning into a wish list of fields.
Step 2: Design the sales process before you touch the platform
Many organisations have sales stages. Few have a sales process. A stage is a label in the CRM. A process gives every stage a purpose, entry criteria, exit criteria, a qualification standard and an owner, so a deal only moves when the evidence says it should.
Here is an illustrative example for a single stage, to adapt to your own process.
Stage element | Illustrative example: "Solution validated" |
|---|---|
Purpose | Confirm the buyer agrees our solution solves a problem worth paying to fix |
Entry criteria | Discovery complete and a business problem documented in the buyer's words |
Exit criteria | The economic buyer has confirmed the problem, the outcome and the budget range |
Evidence required | A meeting note or email from the economic buyer, recorded on the opportunity |
Owner | The account executive, checked by the sales manager in the deal review |
Design stages around how your customers buy, not how your team likes to sell. Buying decisions now involve more people than they used to, so an exit criterion should need agreement across the buying group. One enthusiastic contact is not enough.
This is also where your qualification method belongs. Whether you use MEDDPICC, SPICED or your own approach, the method should show up as evidence standards in the stages. Otherwise it stays in the training deck.
Step 3: Set data standards the pipeline can be trusted on
A pipeline is only as useful as the data in it, and most CRM data is not in good shape. In Validity's 2025 survey of 602 CRM users, 76% said less than half of their organisation's CRM data was accurate and complete. Salesforce's 2024 State of Sales survey found that only 35% of sales professionals completely trust their organisation's data.
Good data standards are mostly about restraint. Four steps cover most of it:
Define ownership. Every dataset and every required field has a named owner.
Specify the fields. For each stage, list the few fields that support a decision. Everything else is optional.
Eliminate duplication. Agree one record for each account, contact and opportunity, and a rule for keeping it that way.
Create one dataset. Sales, marketing and finance report from the same numbers.
Fewer fields, enforced, beat many fields ignored. If a field doesn't support a decision from Step 1, don't make it required.
Step 4: Configure the CRM to enforce the process
Configuration should make the right behaviour the easy behaviour. Stage gates should check for the evidence each stage needs. Required fields should appear only where a decision depends on them. Automation should do the data entry so sellers don't have to.
This matters because sellers already spend most of their week away from selling. Salesforce's 2026 State of Sales survey of 4,050 sales professionals found the average seller spends 40% of their time selling. The same survey found that 42% of reps feel overwhelmed by too many tools. A CRM that adds steps makes both problems worse.
Two questions come up in almost every project:
Do we need a new CRM? Usually not. Most pipeline and forecasting problems come from the process and the data rather than the platform. The same design works in Salesforce, Microsoft Dynamics, HubSpot and most other major platforms.
How long does it take? For the process and CRM design work, two to three months is typical. The pace usually depends on how quickly teams agree definitions.
Step 5: Make adoption part of how the team is managed
Adoption comes from how the team is managed more than from training. Teams keep the CRM current when the people who manage them use it to run the business: pipeline reviews from the CRM, deal reviews against the exit criteria and a forecast call built on CRM data rather than a side spreadsheet.
Gartner's 2020 research made the same point from the other direction. It listed CRM adoption and discipline among the main drivers of forecast accuracy, and found that stronger teams built a culture of following the process.
In practice, three things make adoption stick:
Managers inspect the evidence. A deal review asks "what's the evidence for this stage?" instead of "have you updated the record?"
Missing data has a visible consequence. A deal without its exit evidence doesn't go in the forecast.
The CRM gives something back. Sellers see their own pipeline health, upcoming renewals and next steps without building reports by hand.
How to tell if your CRM transformation is working
Usage measures such as logins and field completion tell you whether people are using the CRM. They don't tell you whether it's helping you sell, so measure the commercial result.
Track a small set of commercial measures before and after the change:
Stage-to-stage conversion, so you can see where deals stall.
Sales cycle length for won and lost deals.
Win rate by segment.
Forecast accuracy against the actual result, and how often close dates slip.
These are the same drivers behind sales velocity. If you want to see how they combine into revenue, our guide to the sales velocity formula works through an example, and Sales Velocity Analysis™ shows which driver is holding revenue back.
Where AI fits in a CRM transformation
AI is only as good as the process behind it. Forecasting tools, deal scoring and AI agents all learn from your CRM data. If stages are inconsistent, fields are incomplete and duplicates are common, AI learns those habits and repeats them at speed.
Salesforce's 2026 survey found that teams using AI agents name manual errors, duplicate data and incomplete data among their biggest data problems. AI comes after the process and the data. Get Steps 2 to 4 right and it becomes useful; skip them and it adds another layer of numbers nobody trusts.
Before you sign off a CRM project, answer these six questions
Which leadership decisions must the CRM support, and who agreed them?
Does every stage have a purpose, entry criteria, exit criteria and an evidence standard?
Which fields are required at each stage, and which decision does each one support?
Who owns the data, and what's the rule for duplicates?
Will pipeline reviews and the forecast call run from the CRM from day one?
Which commercial measures will tell you in six months whether this worked?
If you can't answer them yet, the project isn't ready to configure.
Frequently asked questions
How is CRM transformation different from a CRM implementation?
A CRM implementation changes the tool. A CRM transformation changes how the organisation sells, then makes the tool enforce it. Implementation is part of a transformation, but on its own it rarely fixes forecasting or pipeline quality.
Do we need a new CRM to fix our pipeline and forecasting?
Usually not. Most pipeline and forecasting problems come from undefined stages, weak evidence standards and poor data, and those travel with you to a new platform. Fix the process and data first, then decide whether the platform is really the constraint.
How long does a CRM transformation take?
For the sales process and CRM design work, two to three months is typical. The biggest factor is how quickly sales, marketing and finance agree definitions. Larger data migrations or integrations add time to the build.
Who should own a CRM transformation: sales, revenue operations or IT?
Sales should own the process, the stage definitions and the management rhythm. Revenue operations or IT should own configuration, integration and data. When IT owns the whole project, the platform tends to design the sales process by default.
What percentage of CRM projects fail?
There's no reliable current figure. The often-quoted 70% can't be traced to a source. Gartner's 2001 finding was that 55% of projects failed to meet expectations, which usually means expectations were never defined. Define success in commercial terms before you start, and you can measure it.
Sources
Salesforce, State of Sales, 7th edition, 2026. Survey of 4,050 sales professionals, self-reported. https://www.salesforce.com/news/stories/state-of-sales-report-announcement-2026/
Salesforce, State of Sales report PDF, 7th edition, 2026. https://www.salesforce.com/en-us/wp-content/uploads/sites/4/documents/reports/sales/salesforce-state-of-sales-report-2026.pdf
Salesforce, State of Sales, 6th edition, 2024. https://www.salesforce.com/news/stories/sales-ai-statistics-2024/
Validity, The State of CRM Data Management in 2025. Survey of 602 CRM users, self-reported. https://validity.com/resource-center/the-state-of-crm-data-management-in-2025
Gartner, State of Sales Operations Survey, press release, 12 February 2020. https://www.gartner.com/en/newsroom/press-releases/2020-02-12-gartner-says-less-than-50--of-sales-leaders-and-selle
Rigby, Reichheld and Schefter, "Avoid the Four Perils of CRM", Harvard Business Review, February 2002. https://www.bain.com/insights/avoid-the-four-perils-of-crm/
Bob Thompson, "CRM failure highly exaggerated", interview with Gartner's Ed Thompson, CustomerThink, 2004. https://customerthink.com/reports_crm_failure_highly_exaggerated/
Stewart Rogers, "70 percent of CRM installs fail and other stats you should ignore", VentureBeat, 2015. https://venturebeat.com/technology/70-percent-of-crm-installs-fail-and-other-crap-stats-you-should-ignore
NEXT STEP
See where your sales process breaks
We redesign the sales process your CRM was meant to support, then make the platform enforce it. It starts with a review of your stages, your data and your forecast.
Book A Sales Process Review
WRITTEN BY
Chris Thomas