Replacing a CRM, a marketing platform or a quoting tool takes months. It costs licence fees, migration work and the team's attention. If the real cause is a handover nobody owns, the same delay appears in the new system after go-live.
Find the cause first. You need the people who do the work, a spreadsheet and access to your current system. For one team and one problem, an example schedule is two weeks.
Look for signs in the process and the system
Check the process and the system together, because most reviews find some of each. Confirm each sign with real records.
These signs point to the process:
- Different people give different answers to the same question. Ask three managers how many open deals there are. If you get three numbers, the definitions are unclear.
- Work waits for someone to notice it. Leads sit unassigned. Approved quotes wait for an invoice. Nobody is told it is their turn.
- The last system change did not help. If a previous migration or new feature left the problem in place, look for the cause elsewhere.
These signs point to the system:
- Common tasks take too many steps. People skip a slow screen or field.
- People cannot see or edit what they need. Permissions block a step, so someone else has to do it for them.
- The system cannot record something the agreed process needs. For example, it has no way to track a contract renewal date.
- Data from another system fails or arrives late. An integration drops records, creates duplicates or runs hours behind.
Separate spreadsheets can point either way. Ask the team why they keep them. Distrust of the data points to the process. Slow screens or missing fields point to the system.
Trace one real example from start to finish
Pick one outcome that is going wrong. Good examples are slow replies to new enquiries, deals that stall after a demo, or invoices that go out late. Then choose one recent, real record and follow it through every step.
For each step, write down who did it, which system they used and when it happened. Use timestamps from email, the CRM and calendars. Ask the people involved to fill the gaps. Then work out how long the record waited before each step.
Example (illustrative): a new enquiry at a business-to-business services company. The names are roles.
| Step | Who | Where | Timestamp | Waiting time before this step |
|---|---|---|---|---|
| Enquiry form submitted | Prospect | Website | Monday 9:10 am | Start |
| Form lands in shared inbox | Nobody assigned | Monday 9:10 am | Under 1 minute | |
| Marketing coordinator copies it into the CRM | Marketing | CRM | Tuesday 11:30 am | 26 hours 20 minutes |
| Lead assigned at the weekly allocation meeting | Sales manager | CRM | Thursday 10:00 am | 46 hours 30 minutes |
| Rep calls the prospect | Sales rep | Phone | Thursday 3:40 pm | 5 hours 40 minutes |
The first call happened 78 hours 30 minutes after the enquiry, which is more than three days. The call itself took minutes. Most of the delay came from two waits: an inbox nobody owned and a weekly meeting used to share out leads.
Replacing the CRM alone would leave these waits in place. The team still needs to agree assignment rules and set them up.
Trace three to five records in total. That many can reveal a recurring issue. Check more records before you treat it as typical.
Write down who owns each step
Next, list every step in the process and put one owner beside each. Use a role, such as "Sales manager", so the table survives staff changes.
| Step | Owner | Trigger | Expected time |
|---|---|---|---|
| Assign new enquiry | Sales operations | Form submitted | Within 15 minutes |
| First call | Assigned rep | Lead assigned | Same business day |
| Qualify and record need | Assigned rep | First call held | Within 2 days |
| Send proposal | Assigned rep | Meeting held | Within 3 days |
| Approve discount | Sales manager | Discount over 10% | Within 1 day |
Look for three problems in the finished table. Steps with no owner are where work waits. Steps with two owners are where each person assumes the other has done it. Steps with no trigger depend on someone remembering.
Agree the owner, the trigger and the expected time with the people involved. Then check whether your current system can send the trigger, such as a task or an alert. Check that each owner has the permissions the step needs.
Test whether the team trusts the data
People stop using a system when its data is wrong. Then the data gets worse, because nobody updates it. You can measure this in an afternoon.
- Ask each team for the five fields they use to make decisions. A sales manager might choose deal stage, amount, close date, next step and owner.
- Build a report of recent records in your current system, for example 50. Run it inside the system, and avoid exporting customer data to personal devices.
- For each field, count how many records are blank, out of date or plainly wrong.
- Show the results to the team and ask which numbers match their experience.
Agree with the team how much blank or stale data stops them relying on a report. A common cause is that entering the data gives the person nothing back. A rep updates the close date when a manager uses it in the weekly pipeline review. When nobody uses it, the rep often skips it.
Some gaps come from the system. A field may be hard to find, or an integration may overwrite it. Note these separately, because they need a system change.
A data problem is often a usage problem. Make someone use the field in a decision every week. Training helps once that habit is in place.
Decide what to change
By now you have traced real records, mapped owners and tested the data. Use what you found to choose the change.
| What you found | What to change |
|---|---|
| Work waits between people | Agree owners and triggers. Automate the handover in your current system. |
| Teams define the same thing differently | Agree one definition for each stage and status. Update reports to match. |
| Data is blank or stale | Use the fields in a weekly decision. Remove fields nobody uses. |
| Common tasks take too many steps | Simplify the screens, layouts and required fields. |
| People cannot see or edit what they need | Review roles and permissions with the system administrator. |
| An integration drops, delays or duplicates data | Check its logs, field mapping and error handling. Fix or replace the integration. |
| The system cannot record or show what the agreed process needs | Reconfigure it, or consider replacing it. |
| Workarounds cost more than a replacement over the same period | Plan a replacement, using the written process as the requirement. |
Compare the costs over the same period
Use one period for both sides, for example three years. For a replacement, include licences, migration, configuration, training and the disruption while the team adjusts. For the workaround, include staff time on manual steps, duplicate tools and errors that reach customers or finance.
Replacement makes sense when the agreed process needs something the current system cannot provide, or when the workaround costs more over the period.
Only the last two rows of the table lead to a new system. When they do, the work you have done becomes the requirement. You can give each vendor the process and ask them to show it running in their product.
Run a bounded review
Here is an example schedule for one team and one outcome. Adjust it to the team's availability and the number of records you trace.
- Days 1 to 3: Agree the outcome and how you will measure it, such as hours from enquiry to first call. Trace three to five records and interview the people involved.
- Days 4 and 5: Build the ownership table and run the data test.
- Days 6 and 7: Share the findings with the team. Agree owners, triggers and expected times.
- Days 8 and 9: Check what your current system can do for each change, including permissions and integration logs. Make the small changes.
- Day 10: Decide whether to change the process, reconfigure the system or start a replacement. Set a date to measure the outcome again.