Written by: David Carneal – Digital Efficiency Consulting Group – DECG
Read Time: 3 min
Most companies budget first and design second. That usually produces a solution that fits the budget, not the problem. Then everyone wonders why the results feel underwhelming.
A better approach: design the best process first, then phase it into what you can afford and support.
Why budget-first backfires
- It locks you into current constraints before you’ve explored alternatives.
- It forces teams to optimize the wrong thing (cost) instead of the right thing (flow).
- It encourages half-solutions that create new manual workarounds.
A planning sequence you can steal
- Map the current process (warts included).
- Design the ideal process (no constraints).
- Identify the rules and data requirements (what must be true for the process to work).
- Phase the build (MVP first, then improvements).
- Budget each phase with clear ROI targets.
Phase examples (so this isn’t fluffy)
- Phase 1: Standardize inputs and validations (reduce errors immediately).
- Phase 2: Automate routing and approvals (reduce waiting).
- Phase 3: Integrate systems and auto-generate records (remove retyping).
- Phase 4: Dashboards and alerts (detect issues before they become fires).
A quick ROI sanity check
Estimate time saved per week, multiply by loaded labor cost, then compare to build and maintenance costs. If you can’t explain ROI in one paragraph, the project is probably too big right now.
What to budget for (so nobody gets surprised later)
- Build: design, development, and testing time
- Change management: training and updated SOPs
- Maintenance: rule changes, exception tuning, and monitoring
- Integration: connecting systems and data sources
Your phase 1 should pay for itself
Phase 1 is where you win credibility: reduce errors, shorten cycle time, and remove retyping. Once you have that, later phases are easier to justify.
A one-page decision matrix
Rank each automation idea High/Medium/Low for impact and effort. Start with High impact + Low/Medium effort. Leave the ‘boil the ocean’ projects for later.
- Impact: time saved, error reduction, customer experience
- Effort: complexity, data readiness, integrations required
How to present options to leadership
Leadership doesn’t need the technical details first. They need choices: good, better, best. Each option should include timeline, cost range, risk, and expected impact.
- Good: validations + exception queue (fast ROI)
- Better: add routing/approvals + integrations
- Best: full end-to-end automation + dashboards + alerts
The ‘don’t regret this later’ note
If you automate a bad process, you get a faster bad process. Make sure Phase 0 includes removing unnecessary steps and clarifying ownership. That’s free ROI.
When to stop and re-scope
If the project requires changing five systems, five departments, and five definitions of ‘done’ at the same time, it’s too big. Break it into phases that can ship and prove value.
- Stop signal: nobody can agree on rules
- Fix: run a short mapping workshop and appoint an owner
- Stop signal: data fields mean different things in different teams
- Fix: create a data dictionary and standardize inputs
CTA: Take one workflow and write a 3-line business case: time wasted, error cost, and what ‘better’ looks like. That’s enough to start.