20 — Phase 0 Effort Estimation Analysis¶
Scope: Analysis of time tracking and effort estimates for all 22 Phase 0 epics across M1 (Platform Foundation) and M2 (Shadow Mode Validation). Identifies gaps between current estimates and the late-August 2026 deadline (Week 34).
Deadline: Late August 2026 (~23 August / Week 34). Today: 1 July 2026. Remaining: ~38 working days (7.6 weeks) assuming full attendance. Reduce by 5 days per engineer-week of vacation — July–August is peak holiday season in Sweden.
Data pulled: 2026-07-01 — live Jira (
project = PLT, all Phase 0 epics), including subtask breakdowns and aggregated time tracking.
1. Raw Data — All Phase 0 Epics¶
| Epic | Inc | Summary | Status | Assignee | Orig Est | Remaining | Spent | Subtasks |
|---|---|---|---|---|---|---|---|---|
| PLT-2294 | 1.1 | ID & alias validation | ✅ Closed | Bart De Boer | 2w | 2w | — | 0 |
| PLT-169 | 1.1 | ChangeQueueService | 🟡 In Progress | Johan Ekman | 5w | 1w | 4w | 0 |
| PLT-2792 | 1.1 | Services own CQS queues | 🟡 In Progress | Bart De Boer | Not set | Not set | Not set | 9 |
| PLT-2478 | 1.1 | PS ↔ CQS/DTOflow design | 🟡 In Progress | Sreekanth S. | Not set | Not set | Not set | 0 |
| PLT-1870 | 1.2 | CQS client in R3Server | 🟡 Test | Daniel P. | Not set | Not set | Not set | 0 |
| PLT-2336 | 1.2 | DTOflow PSC accessibility | 🟡 In Progress | Sreekanth S. | 1w | 1w | — | 0 |
| PLT-2118 | 1.2 | DTOflow PROD-ready cert | 🟡 Test | Bart De Boer | 5w | 2d | 4w 3d | 1 |
| PLT-2353 | 2.1 | Pricer Server config export | 🟡 In Progress | Bart De Boer | 1w 1d | 1w 1d | — | 0 |
| PLT-2483 | 2.1 | storeitemvalues export | 🟡 Ready | Johan Ekman | Not set | Not set | Not set | 0 |
| PLT-2496 | 2.1 | Link export | 🟡 Ready | Unassigned | Not set | Not set | Not set | 1 |
| PLT-2494 | 2.1 | ECC params/images/models | 🟡 In Progress | Johan Ekman | Not set | Not set | Not set | 0 |
| PLT-2495 | 2.1 | ECC fonts export | 🟡 Selected | Unassigned | Not set | Not set | Not set | 0 |
| PLT-2492 | 2.1 | ESL Status DTO export | 🟡 In Progress | Bart De Boer | Not set | Not set | Not set | 3 |
| PLT-2488 | 2.1 | itemproperties export | 🟡 Code Review | Unassigned | Not set | Not set | Not set | 2 |
| PLT-2714 | 2.1 | itemproperties startup | 🟡 Selected | Unassigned | Not set | Not set | Not set | 0 |
| PLT-2497 | 2.2 | Consume-ignore-linked mode | ✅ Closed | Unassigned | Not set | Not set | Not set | 0 |
| PLT-2354 | 2.2 | Shadow Mode orchestration | 🟡 In Progress | Daniel P. | 1w 2d | 1w 2d | — | 0 |
| PLT-2359 | 2.2 | ECC Rendering Support | 🟡 In Progress | Bart De Boer | Not set | Not set | Not set | 0 |
| PLT-2966 | 2.3 | API Parity Validation | 🔵 Backlog | Unassigned | Not set | Not set | Not set | 0 |
| PLT-2357 | 2.3 | Linked Item APIs — Items | 🟡 Selected | Unassigned | Not set | Not set | Not set | 0 |
| PLT-2358 | 2.3 | Linked Item APIs — Devices | 🟡 Selected | Unassigned | Not set | Not set | Not set | 0 |
| PLT-2101 | 2.4 | Per-API-path routing | 🟡 Selected | Saikiran K. | 3d | 2d | 1d | 0 |
Aggregate time tracking (sum of epic + all subtask estimates per Jira): PLT-169 has 20w aggregate original (4 subtasks at 5w each), 4w remaining. PLT-2118 has 20w aggregate original, 2d remaining. All other aggregates match the epic-level figure or are not set.
2. Headline Finding: 14 of 22 Epics Have No Estimates¶
| Has Estimate | Count | Epics |
|---|---|---|
| Estimated | 7 | PLT-2294, 169, 2336, 2101, 2118, 2353, 2354 |
| Not set | 15 | PLT-2792, 2478, 1870, 2483, 2496, 2494, 2495, 2492, 2488, 2714, 2497, 2359, 2966, 2357, 2358 |
Of the 11 without estimates: - 2 are closed or Ready for Deploy (PLT-2497, 2483, 2496) — low risk, mostly done - 2 are in Code Review (PLT-2488) or Selected (PLT-2714, 2357, 2358) — implementation mostly done, review or assignment remaining - 2 are In Progress, unestimated (PLT-2792 with 9 subtasks, PLT-2478) — high risk - 2 are In Progress, unestimated, similar to estimated peers (PLT-2494, 2492) — medium risk - 1 is in Test (PLT-1870) — low risk, mostly done
Without estimates, sprint planning is blind. Team capacity cannot be allocated and the lead cannot answer "will Phase 0 close by Week 34" with data.
3. PLT-2354: The Single Most Critical Estimate Error¶
Current estimate: 1w 2d (7 working days)
Why this is wrong: Shadow Mode orchestration is not a feature — it is the integration of every Phase 0 deliverable into one coordinated system. It must:
- Activate parallel execution across all 8 data pipes simultaneously
- Capture and compare every
dtoflow-transmissiondrop against edge output - Validate 100% rendered-image parity across all three Shadow Mode tenants
- Run continuously for 24+ hours without intervention
- Handle failure modes: pipe stalls, image mismatches, config drift, event ordering
Comparable epics for scale reference:
| Epic | What it built | Actual effort |
|---|---|---|
| PLT-169 | ChangeQueueService — single service, single domain | 5w |
| PLT-2118 | DTOflow PROD-ready — single domain certification | 5w |
PLT-2354 must coordinate all domains across all pipes. It is not a 7-day task. A realistic estimate:
| Phase | Activity | Duration |
|---|---|---|
| Skeleton | Connect orchestrator to all pipes, verify each pipe responds | 1w |
| Parallel run | Run all pipes simultaneously, fix ordering/timing issues | 1.5w |
| Parity validation | Compare images, fix rendering diffs | 1.5w |
| Multi-tenant | Repeat on Evo-Se and Application-Stage | 1w |
| Total | ~5w (25 working days) |
Corrected estimate: 5w, not 1w 2d. Delta: +3.5w (+18 days).
This correction alone consumes most of the remaining 38 working days.
4. Suggested Estimates for Unestimated Epics¶
M1 Inc 1.1 — Core Event Routing¶
| Epic | Current | Suggested | Rationale |
|---|---|---|---|
| PLT-2792 | Not set | 2w | 9 subtasks (PLT-2793–2801). Services managing their own CQS queues is configuration + testing work across multiple services. Bart De Boer. |
| PLT-2478 | Not set | 1.5w | PS ↔ CQS/DTOflow integration design. Sreekanth S. |
M1 Inc 1.2 — Internal Comm & Security¶
| Epic | Current | Suggested | Rationale |
|---|---|---|---|
| PLT-1870 | Not set | 3d remaining | In Test. Implementation done. Remaining work is test pass + bug fixes. |
| PLT-2336 | 1w | 1w remaining | DTOflow PSC accessibility. Already In Progress, Sreekanth S. |
| PLT-2118 | 5w | 2d remaining | DTOflow PROD-ready. In Test, Bart De Boer. Aggregate 20w original, 2d remaining. |
M2 Inc 2.1 — Core Data Tap¶
| Epic | Current | Suggested | Rationale |
|---|---|---|---|
| PLT-2483 | Not set | 1d remaining | Ready for Deploy. Done or nearly done. |
| PLT-2496 | Not set | 1d remaining | Ready for Deploy. Done or nearly done. Also unassigned — needs owner. |
| PLT-2494 | Not set | 3d remaining | In Progress, Johan Ekman. ECC params/images/models. Similar pattern to PLT-2483. |
| PLT-2495 | Not set | 2d | Selected, unassigned. ECC fonts export only — narrow scope. Needs owner. |
| PLT-2492 | Not set | 3d remaining | In Progress, Bart De Boer. 3 subtasks. ESL Status DTO. |
| PLT-2488 | Not set | 1d remaining | Code Review. Implementation done. Review + merge only. Unassigned — needs reviewer assigned. |
| PLT-2714 | Not set | 1d remaining | Code Review. Same as PLT-2488. |
M2 Inc 2.3 — API Parity Validation¶
| Epic | Current | Suggested | Rationale |
|---|---|---|---|
| PLT-2966 | Not set | 2w | New epic for Inc 2.3 gate. API parity validation across item, link, and rendering APIs. Needs scoping. |
| PLT-2357 | Not set | 1w | Linked Item APIs — Items. Selected, unassigned. |
| PLT-2358 | Not set | 1w | Linked Item APIs — Devices. Selected, unassigned. |
5. Remaining Effort — Corrected Summary¶
Known remaining (epics with estimates, corrected)¶
| Epic | Status | Corrected Remaining | Assignee |
|---|---|---|---|
| PLT-169 | In Progress | 1w | Johan Ekman |
| PLT-2792 | In Progress | 2w (suggested) | Bart De Boer |
| PLT-2478 | In Progress | 1.5w (suggested) | Sreekanth S. |
| PLT-1870 | Test | 3d (suggested) | Daniel P. |
| PLT-2336 | In Progress | 1w | Sreekanth S. |
| PLT-2101 | Selected | 2d (+ reassign overhead) | Saikiran on vacation |
| PLT-2118 | Test | 2d | Bart De Boer |
| PLT-2353 | In Progress | 1w 1d | Bart De Boer |
| PLT-2483 | Ready | 1d | Johan Ekman |
| PLT-2496 | Ready | 1d | Unassigned |
| PLT-2494 | In Progress | 3d | Johan Ekman |
| PLT-2495 | Selected | 2d | Unassigned |
| PLT-2492 | In Progress | 3d | Bart De Boer |
| PLT-2488 | Code Review | 1d | Unassigned |
| PLT-2714 | Code Review | 1d | Unassigned |
| PLT-2354 | In Progress | 5w (corrected from 1w 2d) | Daniel P. |
| PLT-2359 | In Progress | 1.5w (suggested) | Bart De Boer |
| PLT-2966 | Backlog | 2w (suggested) | Unassigned |
| PLT-2357 | Selected | 1w (suggested) | Unassigned |
| PLT-2358 | Selected | 1w (suggested) | Unassigned |
Totals¶
| Metric | Value |
|---|---|
| Sum of all corrected remaining estimates (person-days) | ~97 person-days |
| Parallelism factor (5 engineers working concurrently) | ÷ 2–3× |
| Critical-path elapsed time (Daniel P. → PLT-2354 dominates) | ~28 working days |
| Available working days (1 Jul → 23 Aug W34, full attendance) | ~38 working days |
| Buffer | ~10 days |
Methodology: Raw person-day totals are not useful for deadline assessment because 5 engineers work in parallel. The critical path runs through Daniel Pettersson (PLT-2354 at 5w = 25d, plus PLT-1870 at 3d, plus Inc 2.2 validation gate at ~5d) = ~33 days. With aggressive overlap between PLT-2354 completion and validation start, the realistic elapsed minimum is ~28 days. This leaves ~10 days of buffer against the W34 target — tight but achievable with focused execution.
6. Per-Person Capacity Analysis¶
| Person | Epics | Corrected Remaining | Feasible in 32 days? |
|---|---|---|---|
| Bart De Boer | PLT-2792 (2w), 2118 (2d), 2353 (1w 1d), 2492 (3d), 2359 (est. 1.5w) | ~5.6w (~28d) | Tight. PLT-2359 adds Phase 0 scope not originally planned. |
| Johan Ekman | PLT-169 (1w), 2483 (1d), 2494 (3d) | ~2w (~10d) | Comfortable |
| Daniel Pettersson | PLT-1870 (3d), 2354 (5w corrected) | ~5.6w (~28d) | Critical path. 28 of 38 available days. 10 days buffer. Manageable if focused. |
| Sreekanth S. | PLT-2478 (1.5w), 2336 (1w) | ~2.5w (~13d) | Comfortable |
| Saikiran Katta | PLT-2101 (2d, now in Inc 2.4) | 2d + reassign overhead | On vacation. PLT-2101 must be reassigned immediately — now on the Phase 0 critical path in M2. |
| Unassigned | PLT-2496 (1d), 2495 (2d), 2488 (1d), 2714 (1d), 2966 (2w), 2357 (1w), 2358 (1w) | ~24d total | Critical. Inc 2.3 has 3 unassigned epics totalling ~4w. Needs owners this week. |
Bart De Boer has the most load — 4 epics plus he's building PLT-2359 (Phase 2 ECC parity) in parallel. This workload is unsustainable for one person.
Daniel Pettersson is on the critical path. PLT-2354 alone consumes 28 of 38 available days. This leaves ~10 days buffer — manageable but requires disciplined execution and no additional scope creep.
7. Can Phase 0 Close by Week 34?¶
Verdict: Achievable with disciplined execution. ~10 days buffer on the critical path.¶
Three factors require active management:
-
PLT-2354 corrected to 5w — Daniel Pettersson's critical path consumes 28 of 38 available days. The remaining ~10 days provide buffer for the validation gates in Inc 2.2 (Shadow Mode Completion) and Inc 2.3 (API Parity Validation).
-
Unassigned epics are invisible work. PLT-2496, 2495, 2488, 2714 — four epics with no owners. Combined they represent ~5 days of work plus review cycles. Until owners are assigned, these don't move.
-
PLT-2101 (Routing, Inc 2.4) is now on the Phase 0 critical path. Saikiran is on vacation. This must be reassigned immediately — it gates the M2 exit.
Realistic timeline with corrections (target end weeks)¶
| Week | Date | Milestone / Increment Target |
|---|---|---|
| W28 | Jul 7-11 | Assign owners. Reassign PLT-2101. Set estimates. Initiate Inc 2.1. |
| W29 | Jul 14-18 | Inc 1.1 target. PLT-169 finishes. PLT-2792, 2478 close. |
| W30 | Jul 21-25 | Inc 2.1 target. Export pipes (2483, 2494, 2496, 2495) close. PLT-2354 skeleton phase begins. |
| W31 | Jul 28-Aug 1 | Inc 1.2 + M1 target. PLT-1870, 2336, 2118 close. PLT-2354 parallel run + parity validation phase. |
| W32 | Aug 4-8 | Inc 2.2 target. PLT-2354 multi-tenant phase + ECC rendering support (PLT-2359). Begin Inc 2.3. |
| W33 | Aug 11-15 | Inc 2.3 target. API parity validation. Buffer week for parity re-runs. |
| W34 | Aug 17-23 | Inc 2.4 + M2 + Phase 0 target. PLT-2101 routing. Final sign-off. |
8. Immediate Actions Required¶
This Week (Week 27)¶
| # | Action | Owner | Impact |
|---|---|---|---|
| 1 | Correct PLT-2354 estimate to 5w in Jira | Daniel P. / Lead | Stops the program from operating on a 1w 2d fiction |
| 2 | Set estimates on all 11 unestimated epics | Lead / epic owners | Enables sprint planning. Use the suggestions in §4 as starting points. |
| 3 | Assign owners to unassigned epics (PLT-2496, 2495, 2488, 2714) | Lead | 5 days of work sitting idle |
| 4 | Reassign PLT-2101 — Saikiran is on vacation | Lead | 2d task. Sreekanth S. or someone with ingress/routing familiarity. |
| 5 | Relieve Bart De Boer — he owns 4 active epics + Phase 2 work | Lead | Spread PLT-2492 or PLT-2353 to another engineer |
| 6 | Assign owners to Inc 2.3 epics (PLT-2966, 2357, 2358) | Lead | 4w of API parity work; PLT-2966 is new and needs scoping. PLT-2357/2358 are Selected but unassigned. |
Before Week 30¶
| # | Action |
|---|---|
| 7 | Run first 24h parity check on Replatforming-Dev as soon as PLT-2354 is skeleton-complete (Inc 2.2). Don't wait for all pipes to be perfect. |
| 8 | Confirm summer vacation schedules for all 5 engineers. Any additional time off in August further compresses the window. |
| 9 | If PLT-2354 is at risk of slipping past Week 32, consider parallelizing Inc 2.3 (API parity) with Inc 2.2 (Shadow Mode Completion) validation. |
9. Risks Introduced by Estimate Gaps¶
| Risk | Severity | Detail |
|---|---|---|
| Deadline set against fictional data | 🔴 Critical | The original mid-August target was agreed with PLT-2354 at 1w 2d. The corrected 5w estimate pushed the realistic close to Week 34. Stakeholders have been informed; the programme now targets Week 34 as the Phase 0 close. |
| Bart De Boer overload | 🟡 High | 6 epics across M1, M2, and M5. No single engineer can sustain this breadth. Risk of burnout + quality degradation. |
| Invisible validation gate | 🟡 High | Inc 2.3 now has three Jira epics (PLT-2966, 2357, 2358) but all are unassigned with no estimates. PLT-2966 is brand new — needs scoping before work can begin. |
| Summer vacation compression | 🔴 High | July–August is peak vacation season in Sweden. 32 working days assumes 100% attendance — an unrealistic assumption. Each week of vacation for Daniel P. or Bart adds 5 days to the critical path. Confirm vacation schedules immediately and rebaseline the timeline. |
| No buffer anywhere | 🟡 Medium | The corrected timeline has zero slack. Any bug, any parity failure, any sick day pushes the deadline. |
10. Recommendations for Going Forward¶
-
Treat PLT-2354 as the programme's heartbeat. It is not "an epic in M2 Inc 2.2" — it is the integration of everything Phase 0 produces. Give it the visibility and resourcing of a Milestone gate, not a feature ticket.
-
Make estimates mandatory before sprint planning. 11 of 18 epics with no estimates is not a data gap — it's a planning failure. Every epic entering a sprint must have an original estimate. Retros should flag unestimated epics as process debt.
-
Scope and assign Inc 2.3 epics. PLT-2966 (new), PLT-2357, and PLT-2358 now exist in Jira for API Parity Validation. PLT-2966 needs immediate scoping. All three need owners and estimates.
-
Re-baseline the Phase 0 target date. With corrected estimates, the realistic close is late August (Week 34). Communicate this to stakeholders now, not in the first week of August when the slip is already visible.
-
Track burndown against corrected estimates. Once all 18 epics have estimates, create a burndown chart from Week 27 to the target close. Update weekly in the status doc.
Companion docs: 15 — Overall Status (current epic states) · 19 — Delivery Framework (Phase → Milestone → Increment → Epic hierarchy) · 17 — Phase 1 Plan (what Phase 0 gates unlock)