Five ways Salesforce implementations fail (andthe fixes)
Fields nobody owns, integrations deferred to phase two, activity metrics as targets — all preventable, none in the sales deck.

Salesforce rarely fails because the software is wrong. It fails because the implementation treated a sales process problem as a software purchase. These are the five failure modes we see repeatedly — all of them preventable, none of them in the sales deck.
1. Automating a process nobody wrote down
The discovery phase produces a diagram of how sales should work. The implementation encodes it. Six weeks after go-live, reps are back in spreadsheets because the actual process — the exceptions, the side channels, the deals that do not fit the stages — was never captured.
Fix: shadow three reps for a week before configuring anything. The real process is visible only in what people do, never in what the playbook says.
2. Fields nobody owns
Every stakeholder gets their required field. Within a year the opportunity form has forty fields, reps fill twelve, and reporting runs on data nobody trusts. This is the most common Salesforce failure and the most boring to fix: delete fields until the form fits on one screen, and give every remaining field a named owner.
A CRM with forty required fields has zero required fields — because nobody fills in any of them accurately.
3. Integration as an afterthought
Salesforce goes live disconnected from the ERP, the support desk and finance. Reps retype order history, support cannot see entitlements, and finance reconciles by hand. The system is blamed for being "just a database" when it was deployed as one.
Fix: the integration list is part of the initial scope, not phase two. Phase two is where integrations go to be quietly cancelled.
4. Admin by committee
Nobody owns the org, so everybody with admin rights makes "just a small change". Two years later there are four hundred workflow rules, half of them contradictory, and nobody can say what happens when an opportunity closes.
Fix: one named admin, a change log, and a rule that automation over a threshold goes through review. Boring governance beats exciting archaeology.
5. Measuring activity instead of outcomes
Dashboards count calls, emails and meetings because those are easy to count. Reps optimise for the count. Pipeline quality falls while activity rises, and leadership concludes the team needs more tooling.
Fix: report on stage conversion and forecast accuracy — the numbers that survive contact with a quarterly review — and treat activity metrics as diagnostic, not targets.
The pattern underneath
Four of the five are process failures wearing software costumes. That is why our Salesforce work starts with the sales floor rather than the setup menu, and why we will sometimes tell a client their problem needs a lighter CRM, or no new software at all. A smaller honest scope beats a large hopeful one — every time we have measured it.
This is the theory. Here is the practice.
The same thinking, applied to a real engagement with measured outcomes.