The real cost of low CRM adoption on forecast accuracy
The pipeline looks full until someone asks why half of it doesn’t close on schedule.
Highlights
- Nearly 40 percent of technology spend underperformed last year because people couldn’t or didn’t use the tools they were given, according to WalkMe’s State of Digital Adoption 2026 report, and CRM systems are one of the most common places that shows up.
- The report traces the problem to a specific moment: a sales rep working across CRM, email, and CPQ, reconstructing deal context by hand because the systems don’t share it automatically.
- Salesforce’s own State of Sales report backs this up directly: 46 percent of sales pros using AI agents say data quality issues are hurting their sales, with manual errors and duplicate data topping the list of problems.
- Organizations closing this gap aren’t running more forecast reviews. Kaseya cut time spent on early-stage opportunities by 45 percent, and Splunk automated its Salesforce quarter-end deal process, saving more than 26,800 hours a year.
Ask a sales leader how confident they are in this quarter’s forecast, and the honest answer is usually somewhere between “mostly” and “depends which rep you ask.” That uncertainty rarely comes from a bad forecasting model. It comes from the data underneath it, and the data underneath it comes from a CRM that reps only partially trust enough to use consistently.
Sales forecasting is only as accurate as the CRM data feeding it, and CRM data is only as accurate as the process reps actually follow when they’re heads-down closing deals, not filling out fields.
Where CRM data quality actually breaks down
WalkMe’s State of Digital Adoption 2026 report describes this problem through a familiar scenario: a sales rep trying to close a deal, working across their CRM, email, and a CPQ tool, with AI starting from scratch at every step because it can’t see the email thread, the call notes, or what was agreed in the last meeting. It doesn’t know what discount the rep already offered. It can’t tell whether the rep is authorized to apply certain terms. And it can’t anticipate what the next person in the approval process actually needs, so the rep ends up piecing everything together manually.
That manual reconstruction is where forecast accuracy quietly erodes. A stage gets updated a week late. A discount gets logged after the fact, if it gets logged at all. A close date gets set to “end of quarter” because nobody has time to confirm the real one. None of that looks like a data quality problem from the outside. It looks like a rep who’s busy, which every sales leader already assumes.
Salesforce’s own State of Sales report puts a number on it: 46 percent of sales pros using AI agents say data quality issues are actively hurting their sales, and when asked what’s driving that, manual errors and duplicate data top the list, ahead of security concerns and incomplete records. The same report finds that 40 percent of sales pros cite forecasting revenue as one of their toughest challenges, even among teams using more sophisticated revenue models built to make forecasting easier. The pattern holds regardless of how advanced the model gets: if the inputs are unreliable, so is everything built on top of them.
“I think the biggest impact of WalkMe is the efficiency it brings.” — Dario Spina, Senior GTM Strategy Analyst, Kaseya
CRM adoption vs. CRM access
Most organizations don’t have a CRM access problem. Reps have logins, the fields exist, and the reports are built. What’s missing is consistent adoption of the process behind those fields, and the two aren’t the same thing.
| Manual CRM entry | Guided CRM entry | |
|---|---|---|
| Deal context | Reconstructed from memory, email, and notes | Captured at the moment the rep is already in the record |
| Required fields | Filled in late, or with placeholder values | Prompted in sequence, before the rep can skip ahead |
| Data quality checks | Caught in a forecast review, after the fact | Caught in real time, before the record is saved |
| Forecast confidence | Varies by rep and by manager’s trust in them | Consistent, because the underlying data is consistent |
The gap between those two columns is rarely about willingness. It’s about whether the system makes the accurate path the easy path.
CRM data quality in practice: two organizations closing the gap
Kaseya: 45% reduction in time spent on early-stage opportunities
Kaseya’s global sales team manages a high volume of deals inside Salesforce, and early opportunity creation, the stage where reps log the deal details that eventually roll up into the forecast, was inconsistent. Without clear guidance in the moment, that data entry got delayed, which made early pipeline visibility harder to trust. The team built real-time, in-app guidance directly into Salesforce, including templates that pre-fill a portion of the required fields, so reps could complete opportunity records accurately without leaving their workflow. The result was a 45 percent reduction in time spent on early-stage opportunities and more consistent data entry across the pipeline, which gave sales leaders cleaner records to build a forecast on and made pipeline reviews and coaching more productive.
Splunk: 26,800+ hours saved annually through automated deal support
Splunk’s quarter-end deal support process used to run through as many as 75 Slack messages per deal, a manual back-and-forth between sellers and deal operations that left plenty of room for details to get lost or logged inconsistently. The team built a Salesforce integration that automatically collects deal information and routes it to the right Slack channel the moment a rep clicks a button, replacing the manual thread with a consistent, structured handoff. That single change saves more than 26,800 hours a year and keeps deal data moving through the pipeline in a form deal operations and finance can actually trust, rather than reconstructing it from a chat log after the fact.
What to look for when closing this gap
A few questions worth asking before assuming the fix is more training or a stricter forecast review process:
- Are required fields getting filled in accurately at the moment of entry, or corrected later during a review?
- Does guidance show up inside Salesforce itself, or does it live in a separate playbook reps have to remember to check?
- Can reps complete early-stage opportunity data without leaving the deal they’re actively working?
- Is there visibility into where in the process reps are dropping off or entering inconsistent data, or does that only surface once it’s already affected the forecast?
- Would fixing the data entry process change how much your sales leaders trust the forecast in front of them today?
The bottom line
Forecast accuracy problems get treated as forecasting problems, when the actual fix usually sits one layer down, in whether reps can enter accurate data without it costing them time they don’t have. Kaseya and Splunk didn’t solve this by asking reps to care more. They made the accurate path through Salesforce the fastest one available.
WalkMe helps organizations close that gap directly inside Salesforce, guiding reps through the fields and workflows that actually feed the forecast, so sales leaders are working from data they can trust the first time, not after a review catches what slipped through.
See how WalkMe supports Salesforce adoption
Frequently Asked Questions
Forecasts are built from CRM fields like stage, close date, and deal value. When reps update those fields late, inconsistently, or with placeholder values because entering them accurately takes too long, the forecast built on top of that data inherits the same inconsistency, regardless of how sound the forecasting model itself is.
Usually process. Most reps already know how the CRM is supposed to be used. The issue is that entering accurate data often takes longer or requires more context-switching than the task warrants, so the path of least resistance becomes an incomplete or delayed entry.
Common indicators include how often required fields are completed on time versus corrected later, how much deal stages shift right before a forecast call, and how much variance exists between what different reps report for similar deal types. A rising gap in any of these usually points back to the entry process, not the reps themselves.
Guiding reps through the specific fields and steps that feed the forecast, at the moment they’re already in the record, tends to move faster than a broader retraining effort. Both Kaseya and Splunk saw measurable results by targeting the exact workflows where data was breaking down, rather than rebuilding their CRM processes from scratch.