ROI tracking
validate the 104-hour/year claim
The workshop's pitch — "2 hours/week per executive = 104 hours/year saved" — has to be validated by actual numbers. Three 3-minute micro-surveys at T+14 / T+30 / T+60 put you in position to present a data-backed ROI readout at the 3-month mark.
§1 Why ROI measurement matters
The "continue / expand" decision on Claude adoption needs to be anchored in measured impact, not estimated impact. Enthusiasm runs high right after the workshop but cools sharply in 2–3 weeks. What actually survives a budget review is concrete numbers: "How many hours saved per person?", "What does that translate to in money?". Estimate-based ROI falls apart at the first CFO question.
The other risk to avoid is the "we did a workshop" trap — the fact of holding the session becomes the only thing on the record, while actual usage stays low. A common failure pattern across IT, HR, and strategy functions. Measure small, chase the root cause if the number is low, use the number as fuel if it's high. Either direction, you've built yourself a briefing deck in 60 days.
§2 Three checkpoints
| When | Purpose | Time | Cadence |
|---|---|---|---|
| T+14 days | Ignition check — did they even try, early-stumble detection | 3 min | 14 days later, start of week |
| T+30 days | Habit check — how often per week, any real time saved | 3 min | 30 days later |
| T+60 days | Initial ROI lock — 60-day cumulative hours saved, cost vs impact | 5 min | 60 days later |
Delivery: Slack DM (bot) or a Google Forms URL. 11 minutes total per executive covers the entire 60-day window.
§3 T+14 survey (3 min)
Theme: did they start, catch early stumbles.
Q1. Times you've opened Claude since the workshop (single select)
- ( ) 0 times
- ( ) 1–2 times
- ( ) 3–5 times
- ( ) 6 times or more
Q2. What you tried (multi-select)
- [ ] Generated something in Code Desktop
- [ ] Used Design to draft a document
- [ ] Tried the Cowork tab
- [ ] Started the 14-Day Learning Path
- [ ] Haven't done anything yet
Q3. Where you got stuck (free text, 2 lines, optional)
______________________________________________
______________________________________________
Q4. What additional support would help (multi-select)
- [ ] Technical hands-on (re-run)
- [ ] Specific prompt examples
- [ ] Internal rollout consultation
- [ ] None for now
§4 T+30 survey (3 min)
Theme: is it becoming a habit, self-estimate of time saved.
Q1. Weeks you used Claude in the last 30 days (0–4 slider)
0 weeks ──●──────── 4 weeks
Q2. Average hours per week (hours/week, number)
[ ____ ] hours / week
Q3. Self-estimated monthly time saved (hours, number)
[ ____ ] hours
Q4. Top 3 use cases you've actually used (multi-select or free text)
- [ ] Document drafting
- [ ] Contract / financial-statement summary
- [ ] Weekly brief auto-generation
- [ ] Code / prototype generation
- [ ] Decision sparring
- [ ] Summarizing internal discussion (Slack etc.)
- [ ] Other: ______________
Q5. Ideas for spreading this in the organization (free text, 3 lines, optional)
______________________________________________
______________________________________________
______________________________________________
§5 T+60 survey (5 min)
Theme: lock in 60-day ROI, extract the next action.
Q1. 60-day cumulative usage time (hours)
[ ____ ] hours
Q2. 60-day cumulative time saved (hours, self-estimated)
[ ____ ] hours
Q3. Monetary conversion (hourly rate × hours saved — use your company's internal rate, typical range ¥8,000–15,000/h)
Rate [ ______ ] yen × saved [ ____ ] hours = [ ________ ] yen
Q4. Top 3 most impactful use cases (free text)
1. ____________________________________________
2. ____________________________________________
3. ____________________________________________
Q5. Continuation decision (single select + reason)
- ( ) Expand
- ( ) Maintain
- ( ) Reduce
- ( ) Stop
Reason:
______________________________________________
Q6. Priorities for the next rollout (multi-select, rankable)
- [ ] Company-wide employee rollout
- [ ] Staged expansion to other teams
- [ ] Switch to Enterprise contract
- [ ] Expand the learning program
- [ ] Other: ______________
Q7. Concerns / issues (free text)
______________________________________________
______________________________________________
Q8. Recommendation to other executives (NPS) (0–10)
0 ─ 1 ─ 2 ─ 3 ─ 4 ─ 5 ─ 6 ─ 7 ─ 8 ─ 9 ─ 10
would not recommend strongly recommend
§6 Dashboard design
At T+60, lay out these eight tiles on one screen. Abstract them to a level that won't get sent back at the board meeting.
- Tile 1 · Avg usage hours / person / week — time series T+30 vs T+60 (habit-forming check)
- Tile 2 · Cumulative time saved — T+60 total (per executive + aggregate)
- Tile 3 · Time saved in yen — hourly rate × cumulative hours saved
- Tile 4 · Continuation intent — expand / maintain / reduce / stop ratio
- Tile 5 · NPS score — T+60 0–10 average with distribution
- Tile 6 · Use-case distribution — T+30 vs T+60 bar chart
- Tile 7 · Stumbling-block trend — T+14 → T+30 → T+60 resolution rate (same barriers lingering?)
- Tile 8 · Sentiment analysis of free text — auto-classified via Claude, positive / neutral / negative ratio
Fastest implementation path: Google Sheets + Looker Studio, or Claude Cowork with a Drive connector regenerating weekly.
§7 Operations
- Distribution: Slack DM (scheduled via Slack bot) or a Google Forms URL in an email.
- Non-response follow-up: One reminder after 3 days. No more than that — respect beats response rate.
- Anonymization: Respondent IDs hashed, no individual identification. Only aggregate numbers go to the board.
- Owner: Corporate Strategy (updates weekly).
- Reporting: One-slide report at T+60 for the board meeting (generate with Claude Design — see §10).
- Retention: Raw data deleted at 180 days; aggregates retained.
§8 Expected ROI ranges
| Scenario | Weekly usage | 60-day cumulative saving | At ¥10,000/hour rate |
|---|---|---|---|
| Conservative | 1 h / person | 20 h | ¥200,000 |
| Base | 2 h / person | 40 h | ¥400,000 |
| Optimistic | 4 h / person | 80 h | ¥800,000 |
Per person, 60 days. Multiply by 3 for three executives. Base scenario annualized is ~240 hours = ¥2.4M / year per person, which comfortably clears the Max plan cost for three executives.
§9 Playbook for negative ROI
If any executive shows "0 usage" at T+30, the right response is not to scold. It's to switch into "remove the barrier" mode. Run this sequence:
- Step 1: One-on-one (15 min) — identify the barrier: technical / time / motivation / no use case.
- Step 2: Re-run hands-on (30 min, Part 1 of the workshop only) — focus only on "can you launch it."
- Step 3: Re-propose use cases — tailored to that executive's specific work, not generic.
- Step 4: If still stuck, honestly park it. Not every tool fits every person — don't force it.
§10 Three-month board readout template
At T+60, present to the board with these 10 slides. Feed the §6 dashboard data into Claude Design and a draft comes out in ~10 minutes.
- Cover: "Claude adoption — 60-day ROI"
- Timeline: 60-day progression from the workshop
- Cumulative usage / time saved / monetary impact: tiles 1–3 on a single slide
- Use-case distribution: tile 6 bar chart
- Top 3 success cases: who (anonymized), what for, what impact
- Barriers and resolution: tile 7 trend + §9 playbook execution status
- NPS score: tile 5, with distribution
- Three options: expand / maintain / reduce — cost and expected impact for each
- Recommendation + next 90-day plan: which to pick, what to start first
- Q&A: 5 anticipated questions + response skeletons
- Feedback survey — day-of satisfaction (different metric)
- 14-Day Learning Path — action design aligned with T+14
- Claude Cowork deep dive — source for reference ROI numbers
- Advanced course — next learning program if ROI is positive