Mission Engineering 07: Baseline First — Test Process Changes Before Buying Gear

Establish a Fixed Comparison Baseline

The author opens with a common dilemma: morning coffee feels rushed, so one might consider buying an electric grinder or adding more detailed pour records. Both ideas seem reasonable, but without knowing whether grinding is the real bottleneck or whether extra recording will consume scarce time, any change may miss the actual problem. The previous article identified who does what and with which tools, leaving a few potential busy spots. This article explores two improvements from those spots — advance preparation versus additional recording — and compares their benefits, costs, and abandonment conditions. It starts by laying out a comparison plan; no experimental results are presented yet.

Lock Down a Comparison Starting Point

If tomorrow you pre‑set equipment, switch beans, and skip a few records, and the coffee arrives earlier, which change caused the improvement? A single run cannot tell. Therefore the first step is to write down the agreed comparison starting point, called a Baseline in Mission Engineering. A baseline provides a reference; it does not claim to be optimal or fully validated.

The draft baseline (CF‑PLAN00) fixes:

Scenario & Goal: Workday scenario CF‑WD; start prep at 07:20, target serve by 07:28, family leaves at 07:35. Taste and recording requirements follow article 05; undecided items to be confirmed.

Initial Conditions: Water not boiled, beans not ground, equipment not placed; materials and countertop checked against scenario card. Actual conditions recorded per run.

People & Equipment: You (from article 06), this kettle, hand grinder, digital scale, phone, filter and server; no new equipment this round.

Procedure: Select one recipe version; write the exact sequence of prep actions and allowed overlaps into an execution plan — not just "do as usual".

Basic Records: Adopt confirmed candidate list from article 05: recipe version, bean batch, actual bean and water weights, start and serve times, ad‑hoc changes.

Comparison Criteria: CF‑S01 taste rating, CF‑S02 on‑time delivery, CF‑T01 morning elapsed time, CF‑E01 prep readiness, CF‑R01 missing records; keep definition versions consistent.

Three plans are numbered: baseline draft CF‑PLAN00, Option A CF‑PLAN01, Option B CF‑PLAN02. PLAN is the series' plan numbering; all are draft v0.1. Same plan can be tried multiple times; each run records who, when, and outcome separately.

Identical equipment does not guarantee identical arrangement. Kettle start order, when to read the scale after weighing — these affect comparison. Therefore the operation sequence, check actions, and recording points left undecided in article 06 must be finalised before trial; a still‑fluid procedure cannot serve as a fixed reference.

State Clearly What Each Option Changes

This round purchases no equipment; it compares two changes, both branching from CF‑PLAN00: Option A moves selected placement to the night before (no pre‑grinding, no pre‑boiling, no extra segment records). Option B keeps baseline prep but adds start/end times and volumes for each pour segment to the basic records. Testing them separately isolates each change's impact. Later they can be combined in a new comparison to see if time saved by advance prep can absorb the recording overhead.

Baseline CF‑PLAN00: Follow confirmed current arrangement; keep basic records. Expected improvement: provides reference for all results. Costs to check: may still have waits, missing items, or awkward ops.

Option A: CF‑PLAN01: Move chosen placement to previous night; no pre‑grind, pre‑boil, or extra segment records. Expected improvement: less morning equipment hunting; possibly earlier serve. Costs to check: extra evening time, countertop occupation, morning re‑checks needed.

Option B: CF‑PLAN02: Keep baseline prep; add per‑segment pour start/end times and volumes to basic records. Expected improvement: easier to reconstruct each segment for future replication. Costs to check: reading scale, switching phone screens, entry delay; can readings be retained, can omissions be recovered.

Option A's "placement" must specify which items, where. Next morning check if moved or still usable; if directly usable, it differs from the original morning hunt. Record this change in the CF‑WD comparison log; do not claim conditions were identical.

Option B does not alter the recipe's segment structure. The example assumes the chosen recipe already has distinguishable pour segments; if not, choose a different recording question rather than invent segments for the table. Whether each segment's volume is single‑pour or cumulative, and how to read it, must be agreed beforehand.

The term "option" is everyday Chinese. The Mission Architecture Style Guide (MASG) distinguishes Alternative (introduces new technology/concept) from Excursion (examines changes to baseline conditions, resources, behaviours, assumptions). Here advance prep and adjusted recording are treated as procedure and assumption changes. Adding a few fields does not yet constitute a new concept; if a whole recording method is swapped later, re‑evaluate what actually changed. Do not categorise merely because "equipment changed" or "weather changed".

Wanting an electric grinder can be noted. This round agreed no purchase; first compare methods with current gear. Later, if purchase is considered, check which manual ops it reduces, whether it adds checks or cleanup, and whether those changes bring coffee to the table earlier — then decide if worth it.

Where Does Option B's Recording Fit in the Gaps?

"Record every segment" is easy to write in a comparison table; the hard part is finding when to record. A candidate arrangement: after finishing a pour and steadying the kettle, read and enter that segment's data before the next segment's operation begins. Two preconditions: required data still retrievable, and the gap long enough to complete entry.

Using resource IDs from article 06, the plan is sketched below. Times are unmeasured; only events are listed, no plausible seconds filled in.

Manual pour for this segment — Resources: CF‑RES01 you, CF‑RES02 kettle; also need scale and timer. Planned occupancy: from picking up kettle to pour end and kettle steady; per article 06 this segment does not overlap manual entry.

Read & enter this segment's data — Resources: CF‑RES01 you, CF‑RES05 phone. Planned occupancy: planned after kettle steady; occupancy from picking up phone, switching screens, through reading and input to save. First verify start time etc. have original evidence; cannot backfill after the fact.

Next segment operation — Resources: still needs CF‑RES01, plus kettle etc. per procedure. Planned occupancy: start per recipe's time or trigger; check if previous entry finished, whether switching back to timer costs time.

The table retains a concrete question: if entry isn't finished when the next pour must start, what to do? Cannot allow simultaneous continuous pouring and typing, nor silently shift the next segment's start time for recording's sake.

Can reduce on‑the‑spot entry, keep raw data and organise later, or use a calmer morning for detailed recording trial. But must verify: does the phone actually save these data? Can they be found later? If tried on a weekend, which results still apply to workdays?

Two question marks are noted beside Option B: "When to record? Will data be retrievable then?" Article 08 will put pour and entry into one model; article 09 checks if they crowd each other. Even if crowded, continue to see if it's wait, omission, or delay — do not label it deadlock prematurely. Real behaviour tested in article 11.

Lay Out Test Questions in a Matrix, Don't Rush to Score Options

First place the procedures to compare, test conditions, and questions to answer in one table. Mission Engineering Guide uses a Run Matrix for this, requiring clear assumptions, constraints, boundaries. In this small case, an Analysis & Test Plan table is used.

Each row lists one question; how many runs, how to record and compare, agreed before start. If conditions change between runs, verify comparability.

Each actual run follows article 02's method: assign a brew ID, note plan version, scenario, recipe. Three runs of same plan yield three records; paper‑only analyses go to analysis log.

Compare Baseline vs Option A: Keep recipe, materials, operator, equipment consistent; A changes advance prep content. Data to capture: morning start/end; evening, morning, post‑cleanup manual times; waits and rework; taste and on‑time. Questions: earlier serve? Less morning rush? Total human effort reduced?

Compare Baseline vs Option B: Keep baseline prep; B adds agreed segment records. Data: CF‑R01 basic missing items, extra segment missing items, entry occupancy and op deviation, plus CF‑S01, CF‑S02, CF‑T01. Questions: what extra reliable info gained? Does it add waits or affect coffee result?

Option B hits one info‑view interruption: In desk walk‑through assume timer screen temporarily invisible; note which action, how long. Data: which data lost, which op affected, how re‑verified after recovery. Question: does current recording arrangement assume "info always visible"?

Re‑do separately per baseline and Option B records: Agree separate comparable conditions, personnel, evaluation methods; use both record sets, note order, prior experience, condition differences. Data: which steps reconstructable from records, which still need follow‑up, plus each run's taste and time. Question: are B's extra records more helpful than basic? Testing B alone cannot conclude it beats baseline.

Row three is first a paper walk‑through; no need to deliberately create interference with hot water. If real trial needed, design separately. Paper discovery of possible data loss ≠ field loss already occurred.

When comparing record completeness, first compare the basic items both plans require, then list Option B's added segment items separately. Different item counts make direct "completeness rate" misleading. B's extra info must be checked for usefulness in next replication.

Time must also be split. CF‑T01 is elapsed from first morning prep to serve, including waits. Human effort sums actual op times per segment, plus evening prep, post‑cleanup, supplementary entry. When kettle heating overlaps hand grinding, cannot add both durations as human effort; nor omit post‑serve recording work.

This table lists what to compare. Article 11 will use baseline and Option B as example to detail timing method, which to try first, and how to watch for learning effects. Option A trial would need same details filled in, otherwise only "faster" is seen without knowing why.

What Result Would Make You Abandon This Change?

Writing the abandonment condition before comparison prevents cherry‑picking favourable numbers afterward.

Option A moves some prep to the night before to ease morning. If trial shows earlier serve but evening + morning + cleanup total time increases, both facts must be recorded. Spending a bit more evening time for a calmer morning may be acceptable to you and family. Weigh the time account together with taste and serve time; don't pick only "morning faster".

If Option B reduces missing records but delays a critical operation or misses the agreed serve time, pause that recording arrangement, diagnose cause, then adjust. More complete records cannot override must‑meet requirements. If requirements truly need change, both parties re‑agree and apply from that run forward — cannot retroactively turn a missed result into success.

Conversely, if B's extra records cost little time yet still fail to explain next replication, re‑examine whether the wrong focus was recorded. Missing one key condition may hurt usability more than missing several duplicate readings.

Writing these expectations as checkable statements is more useful than filling "effect improved, efficiency increased". Can add a Plan Card to existing archives:

Plan ID & Version:
Baseline Version Compared:
Questions to Answer, Linked Requirements & Metrics:
What Changed, What Conditions Held, What Else Shifted:
Expected Improvement & Why It Might Work:
New Actions, Resource Use & Handovers:
How to Analyse or Trial, Where Raw Evidence Stored:
Results Supporting Adoption, Results Requiring Revision or Abandonment:
Unverified Conditions & Conclusions:
Who Confirms Comparison Plan & Subsequent Choice:

Applied to your own business, pick a similar spare‑parts issuance case: compare two logging methods — line‑by‑line entry vs. record during physical check then batch entry. Beyond time saved, watch for wrong associations, omissions, follow‑up questions. Do not delete existing verification steps until trial proves reliability.

Change One Plan — What Else Must Change?

Option B appears to add only a few record fields, but requires tracing back several places: in article 05's requirements, which are basic items and which are trial‑only additions? In article 06's schedule, when are person and phone free? At which step is raw data captured, which organisation tasks can shift to A05? Insert concrete actions into execution plan; if tasks or dependencies are missing, update the task table.

Further down, recipe is information to follow, brewing is the occurring process, kettle and scale provide support, coffee grounds, water, and liquid coffee transform during the process. Adding one row in one table easily leaves old arrangements elsewhere.

With few tables, manual cross‑check works. When pour schedule changes and you must re‑check person, kettle, records, flipping back and forth becomes painful. That's why the next article introduces OPM (Object‑Process Methodology) to hold these relationships in one model. The model must still trace back to original requirements, metrics, records, and resource constraints; after modelling, verify feasibility.

Summary

First secure a clear comparison baseline, then spell out what each change alters, what improvement is hoped for, and how to detect failure. Even if the final decision is to keep current equipment and recording level, there is documented reasoning why.

Next article will model this coffee analysis in OPM to see how processes transform objects and how people and equipment participate.

References

U.S. Department of Defense, Mission Engineering Guide , Version 2.0, 2023‑10‑01, §5.3 (p.22) and §6.1.1‑6.1.2 (pp.24‑26). Used for baseline, alternative change recording, Run Matrix, analysis method selection. This article's plan numbering, comparison tables, recording arrangements, and trial examples are instructional designs; no measured conclusions provided. https://ac.cto.mil/wp-content/uploads/2023/11/MEG_2_Oct2023.pdf

U.S. Department of Defense, Mission Architecture Style Guide , Version 1.0, cover date 2025‑01‑06, Chapter 5, pp.21‑22. This article follows that context to distinguish Baseline, Alternative, and Excursion, not reducing them to fixed equipment or environment categories. https://ac.cto.mil/wp-content/uploads/2025/01/U-Mission-Architecture-Style-Guide-Final_07Jan2025.pdf