Discovery sets the foundation; mistakes here compound through the project.
1. Skipping Discovery to "save time".
Pressure from sponsor to get to Build fast. Result: Build phase repeatedly hits unknowns; cost overrun far exceeds the "saved" Discovery time.
Fix: educate the sponsor on the risk. Cite past projects. Ground the timeline.
2. Talking only to executives.
Executives describe what they think should happen; not what actually happens. Result: build to imagined process, not real one.
Fix: at minimum 2-3 hours per affected role with actual users.
3. Not mapping current state.
Jumping to future state without understanding current state. Result: future state may be unrealistic; transitions painful.
Fix: explicit current-state mapping, even if "it's a mess and we're fixing it" is the conclusion.
4. Assuming standard solutions fit.
"Out of the box Sales Cloud will work" — turns out the customer has a unique sales motion the standard model doesn't capture.
Fix: validate fit during Discovery with concrete scenarios, not abstract descriptions.
5. Underestimating data migration.
"We have data" gets glossed over. Reality: source data is messy, voluminous, and migration is months of effort.
Fix: profile data during Discovery. Sample the source. Identify dupes, gaps, formats.
6. Glossing over integrations.
"We need integration with SAP" — assumed simple. Reality: SAP team's availability + technical complexity = months.
Fix: engage external system owners during Discovery. Validate technical feasibility, schedules, formats.
7. Ignoring change management.
Discovery captures functional requirements but misses adoption considerations. Result: brilliant build, low usage.
Fix: include change management as a workstream in Discovery.
8. Vague output documents.
SDD with "the system should support sales process improvement" rather than concrete details. Result: Build team doesn't know what to build.
Fix: every requirement should be specific, observable, testable. "If a vague requirement survives Discovery, it'll fight you in Build."
9. No risk identification.
Discovery delivers an SDD with no acknowledged risks. Reality: every project has them.
Fix: explicit risk register. Reviewed weekly.
10. Not surfacing political/organizational factors.
Discovery focuses on tech and process; misses that "Bob in Sales is opposed because he built the legacy system."
Fix: stakeholder interviews probe attitudes; document politics where relevant.
A senior consultant treats Discovery as the highest-leverage phase. The reason: it's cheaper to be right early than to fix it later.