1. Building the pipeline first
The tool invites it. You log in, you see a board, you start naming stages. Two hours later the pipeline is done and nobody agreed on anything.
The fix is boring. Write the process on one page before you touch the software. Ten sentences beat a clever board.
2. Stages that describe your feelings
Stages called Interested, Hot and Almost there cannot be checked. Two colleagues will place the same deal in different stages, and the forecast becomes an opinion.
Name stages after observable events. Discovery call held. Requirements confirmed. Proposal sent. Verbal agreement received. Each of those has a yes or no answer.
3. Too many pipelines
Teams often create a pipeline per product, per country or per rep. Reporting then becomes a manual merge, and deals get stuck in the wrong board.
Use fields for product, country and segment. Use a second pipeline only when the steps genuinely differ, for example new business versus renewals.
4. Forty custom fields
Fields feel free. They are not. Every extra field lowers the chance that the important ones get filled.
Start with the reports you need. Keep eight to twelve deal fields. Review after thirty days and delete what stayed empty.
5. Free text where you need a list
Deal source as free text produces Website, website, Web site and Inbound in the same quarter. You cannot filter it, so you cannot learn from it.
Use single option lists for anything you plan to group by. Keep the list short enough that reps read all of it.
6. Importing dirty data
A CRM full of duplicates loses trust in the first week. Once a rep finds two versions of their own customer, they stop believing the system.
Clean in the export file. Deduplicate on email and on domain. Delete records with no activity in three years. Import organisations, then people, then deals, then activities.
7. Migrating everything because you can
Ten years of lost deals do not help anyone. They slow search and clutter reports.
Set a cut off. Open deals, plus won deals from the last two years, plus the contacts attached to them. Archive the rest outside the CRM.
8. Automating an unstable process
Automation locks in whatever you have. If the process changes in month two, the automations fight you, and nobody remembers why a deal suddenly moved.
Automate the safe things first. Follow up activities, idle deal reminders, defaults on creation. Wait a month before automating anything that emails a customer.
9. No internal owner
This is the mistake that kills projects quietly. Without one named person, questions go unanswered, exceptions become habits, and the setup drifts within a quarter.
Name an owner before the project starts. Give them time, not just the title. Two hours a week in the first three months is realistic.
10. Training once, at go live
People forget a demo in three days. They learn by doing their own work in the tool while someone answers questions.
Run a short session at go live, then a second one two weeks later using their real deals. The second session is where adoption is won.
11. Reporting designed last
If reporting is an afterthought, you will discover in month two that the field you need does not exist, and the last hundred deals cannot be corrected.
Define five reports before configuration. Pipeline value by stage. Conversion per stage. Won revenue by source. Activity per rep. Average sales cycle. Then build the fields those need.
12. Treating email sync as optional
If email lives outside the CRM, the CRM is a form to fill in. Reps will not do that for long.
Connect mail and calendar on day one. Agree what stays private. When conversation history is in the deal, the CRM starts saving people time instead of costing it.
What to do if you already made three of these
Do not rebuild from scratch. Rebuilding throws away the habits your team has already formed, and habits are the expensive part.
Fix in this order. Stages first, because reports depend on them. Then fields, then data, then automations. Change one thing per week and tell the team why.
The Readiness Assessment ranks which of these applies to you. The follow up email names the order of repair.
Questions
What is the most common Pipedrive implementation mistake?
Building the pipeline before agreeing the sales process. Stages then describe internal hope instead of buyer behaviour, and every report built on them is unreliable.
How many custom fields should we start with?
Around eight to twelve on deals, and only the ones a planned report needs. Fields are easy to add later and painful to remove once reps have filled them inconsistently.