# Derivative Content Plan
## What Counts as a Change Order? Flagship to content system

**Publisher:** Northline Field Research
**Anchor asset:** *What Counts as a Change Order?*, a ten-page original research brief built from 130,782 public federal contract actions
**Covers:** the full derivative program, the phase-one build order, the launch sequence, and how each tier is measured

The brief is an input to a content program rather than the end of one. This document maps every asset the research can support, sets the order in which they are produced, and defines what each tier is accountable for.

---

## 1. What the playbook says

The practice has a name and a fairly settled shape. Content atomization takes one anchor asset and maps every derivative before production begins, rather than adapting pieces after the fact. The distinction from ordinary repurposing matters: repurposing is reactive, atomization is planned, and the anchor gets designed with fragmentation in mind. Practitioners cite a working ratio of roughly 15 to 30 distinct pieces across six or seven channels from a single 3,000-word research report.

The common operating framework is PADR: pick a **P**illar with original data and a clear point of view, **A**tomize it into channel-ready pieces, **D**istribute where the audience already is, **R**efresh with new data later.

Three failure modes recur in the literature, and all three are avoidable here:

- **Republishing instead of repurposing.** Pasting the same paragraphs into a LinkedIn post fails. Each derivative has to be rebuilt for its format.
- **Treating the launch as an event.** Launch day is the midpoint of a sequence, not the finish line. Sales-facing material should exist before the public announcement, not after.
- **Optimizing for the wrong metric.** Webinar registration counts and press pickups are vanity numbers if nothing downstream connects to them.

One point specific to this paper: original research carries the highest citation value of any content type and is unusually friendly to AI answer engines, because it produces facts that cannot be synthesized from existing material. That makes the dataset and methodology assets in their own right rather than supporting material.

---

## 2. Atom inventory

Everything extractable from the current draft, listed before any format decisions. This is the raw material.

**Findings**
- A1. The headline reframe: 3.24% is an accurate total for one label, not the cost of change orders.
- A2. The audit result: 39 of 80 sampled descriptions clearly changed the required work; 21 changed schedule only.
- A3. The definition gap: 3.24% becomes 17.07% depending on which codes count.
- A4. The size gradient: change-order labels appear on 5.1% of contracts under $100K and 59.3% of contracts over $5M.
- A5. The high-dollar anecdote: a $10.6 million option exercise filed under a change-order code.
- A6. The B-record composition: nine option records worth $658.1 million dominating the largest supplemental agreements.

**Frameworks**
- A7. The Change Evidence Standard, seven fields: Condition, Evidence, Direction, Impact, Classification, Resolution, Closure.
- A8. The three coding lenses (narrow, moderate, broad) as a diagnostic method anyone can apply to their own data.
- A9. The limitations matrix: what the data can and cannot show.

**Narrative**
- A10. The 9:17 Tuesday morning opening: one condition becoming five records.
- A11. The thesis-shift story: setting out to price change orders and discovering a classification problem instead.
- A12. The methodology story: three failed data pulls, the truncation discovery, the progressive-scope contamination.

**Assets already built**
- A13. Nine data graphics plus a demo QR, in PNG and SVG.
- A14. The frozen analysis, manifest, claim ledger, source ledger, audit workbook.
- A15. The extraction and analysis script.

Fifteen atoms is comfortably enough for a 20-plus-asset program.

---

## 3. The full ecosystem

### Tier 1: Companion assets (gated or downloadable)

| # | Asset | Atoms | Notes |
|---|---|---|---|
| 1.1 | **The one-page Change Evidence Standard** | A7, image8 | The paper's CTA already promises this. Until it exists, the flagship has a broken promise on its back cover. Highest priority derivative by a wide margin. |
| 1.2 | **Project self-audit worksheet** | A7, A10 | Trace one recent change from field evidence through authorization, cost review, billing, closeout. Mark where the trail breaks. The CTA promises this too. |
| 1.3 | **Executive summary one-pager** | A1–A4, image2 | Two pages maximum, built to be forwarded by a champion to a CFO who will not read ten pages. |
| 1.4 | **Change classification codebook** | A8 | The taxonomy as a working tool: how to separate time-only, option, administrative, and substantive scope in your own system. The most directly usable thing in the whole project. |
| 1.5 | **Reproducibility package** | A14, A15 | Public repo: script, methodology, frozen files, manifest. Near-complete already. This is the AI-citation and credibility play. |

### Tier 2: Article series

Five posts, each standing alone, each pointing back to the paper. Order matters: lead with the story, not the number.

| # | Working title | Atoms | Why it works |
|---|---|---|---|
| 2.1 | Why I stopped trying to calculate what change orders cost | A11, A1 | The thesis-shift narrative. Best single post in the series: it has a reversal, it teaches a method, and it is the least reproducible by anyone else. |
| 2.2 | One condition, five records | A10, image3 | The narrative section expanded, with the handoff diagram. The most relatable piece for a field audience. |
| 2.3 | What is actually inside a change-order code | A2, A5, image5 | The audit, anchored by the $10.6M option anecdote. |
| 2.4 | Big projects change more often. Small projects change more. | A4, image6 | The counterintuitive finding, and the most quotable single chart. |
| 2.5 | A change record that survives every handoff | A7, image8 | The practical payoff. Ends with the one-pager download. |

A sixth, optional and aimed at a different audience: **How I checked my own numbers** (A12, A9, image4, image9), a process post for people who build research rather than consume it. It reaches analysts and researchers rather than operators, so it is scheduled outside the main series.

### Tier 3: Social

| # | Asset | Atoms | Format notes |
|---|---|---|---|
| 3.1 | Definition-gap carousel | A3, A6, image7 | Eight to ten slides. The strongest carousel candidate because the ladder structure is already sequential. |
| 3.2 | Audit-composition carousel | A2, image2, image5 | "What 80 change-order records actually contained." |
| 3.3 | The $10.6M hook post | A5 | Single anecdote, no chart needed. Highest-scroll-stopping atom in the set. |
| 3.4 | Size-gradient chart post | A4, image6 | Chart plus three sentences. |
| 3.5 | Seven-field standard post | A7, image8 | The framework as a standalone reference graphic people save. |
| 3.6 | Process thread: three failed data pulls | A12 | Behind-the-scenes content. Performs well with practitioner audiences and costs nothing to make. |
| 3.7 | Limitations post | A9, image9 | "Here is what my own research cannot tell you." Counter-programming against confident vendor research. |
| 3.8 | Short video over the charts | A1, A3, A4 | Sixty to ninety seconds, screen recording with voiceover. |

### Tier 4: Live and spoken

| # | Asset | Atoms | Notes |
|---|---|---|---|
| 4.1 | Webinar: *What is actually in your change-order data?* | A1–A3, A7 | Forty minutes: the finding, the audit method, the standard, then live Q&A. Bring a construction attorney or claims consultant as guest so it stops being a monologue. Target a narrow qualified audience rather than maximum registrations. |
| 4.2 | Conference or association talk | A10, A1, A7 | Same spine, twenty minutes, heavier on the narrative opening. |
| 4.3 | Podcast guest kit | A5, A11 | Three pitch angles, prepared talking points, one clean anecdote per angle. Cheap to produce, high reach per hour. |
| 4.4 | Webinar cut-downs | 4.1 output | Three clips of two to three minutes each, plus a recap post and an on-demand gated replay. |

### Tier 5: Sales and account use

Included because a flagship that never reaches a sales conversation is doing half its job. This tier is produced with the sales team, not handed to them.

| # | Asset | Atoms | Notes |
|---|---|---|---|
| 5.1 | Role-specific one-pagers | A1–A4 | Three versions: operations (standardization), finance (what the label conceals), risk (evidence quality). Same research, three framings. |
| 5.2 | Discovery question set | A7 | Seven questions derived from the seven fields. Turns the framework into a conversation guide. |
| 5.3 | Objection handler | A2, A3 | For "we already track our change orders." The audit is the answer. |
| 5.4 | Benchmark conversation starter | A4 | "Companies your size see labels on roughly this share of contracts. What does your data show?" |

### Tier 6: Earned and syndicated

| # | Asset | Atoms | Notes |
|---|---|---|---|
| 6.1 | Trade press pitch | A1, A4, A5 | ENR, Construction Dive, Roofing Contractor, ACHR News. Lead with the definition gap, not the paper. |
| 6.2 | Bylined trade article | A1, A7 | Original 800 words for a trade outlet, not a reprint. |
| 6.3 | Association or newsletter feature | A4, A7 | AGC, ABC, NAHB chapter newsletters. |
| 6.4 | Dataset as citable resource | A14, A15 | Published, documented, linkable. The asset most likely to be cited by other researchers and by AI answer engines. |

That is 30 assets from one ten-page paper, which sits at the upper end of the ratio the practice literature reports.

---

## 4. Specimens

Descriptions of derivatives are easy. The proof is in whether the reshaping is real, so here are three built out.

### 4.1 LinkedIn post, the $10.6M hook (asset 3.3)

> I went looking for what change orders cost the construction industry.
>
> I found a $10.6 million option exercise filed under a change-order code.
>
> The federal contract database is precise about how a transaction was entered, but it cannot tell you what actually happened on the job.
>
> So I read 200 of them by hand. In a random sample of 80 records officially labeled as change orders, 39 clearly changed the required work. Twenty-one only changed the schedule. Twelve were too thin to tell.
>
> The number everyone wants is "what percentage do change orders add?" The number I can defend is much smaller: 3.24% of original commitments, across 1,481 completed federal construction contracts, for records carrying one particular label.
>
> That is not the cost of change. It is the total for one filing category, which encompasses a number of different record types.
>
> Full analysis, method, and the seven-field record standard I built from it: [link]

Note what changed from the paper: first person throughout, one idea per paragraph, the anecdote promoted to the opening, the caveat kept but compressed, and no chart because the story carries it.

### 4.2 Carousel outline, the definition gap (asset 3.1)

Nine slides, sourced from image7, reformatted from landscape to portrait.

1. **Cover.** "Three ways to count the same change orders. Three very different answers."
2. Narrow definition: explicit change-order codes only. **3.24%**
3. Add within-scope supplemental agreements. **17.07%**
4. Add additional-work agreements. **17.36%**
5. "A five-fold swing, from the same contracts, in the same eleven years."
6. What was inside the largest supplemental agreements: nine option records, $658.1 million.
7. "Options are not changes. They were established or triggered under the original agreement."
8. "The spread is not a range of estimates. It is a warning about categories."
9. **CTA.** Full method and the seven-field standard.

### 4.3 Webinar run of show (asset 4.1), 40 minutes

- **0:00** Cold open with the 9:17 Tuesday morning story (A10, image3). No agenda slide.
- **0:04** The question I set out to answer, and why the answer broke (A11).
- **0:10** What the audit found (A2, A5, image5). The $10.6M record as the turn.
- **0:18** The definition gap (A3, image7).
- **0:24** Guest segment: a construction attorney or claims consultant on what documentation actually holds up. Unscripted conversation, since that is where the reusable clips come from.
- **0:32** The seven-field standard (A7, image8) and the self-audit.
- **0:36** Q&A.

Plan the three clips before recording rather than hunting for them afterward: the $10.6M turn, the guest's best answer, and the standard walkthrough.

---

## 5. Image reuse map

All nine data graphics exist at 1800px wide in PNG with SVG masters, so they can be recut without redrawing.

| Image | Original use | Reuse |
|---|---|---|
| image1 | Cover chain | Launch announcement card, article series header |
| image2 | Figure 1, summary | Exec one-pager hero, carousel 3.2, newsletter |
| image3 | Figure 2, handoff chain | Post 2.2, webinar cold open, self-audit worksheet |
| image4 | Figure 3, cohort funnel | Method post, webinar credibility slide, repo README |
| image5 | Figure 4, audit | Post 2.3, carousel 3.2, webinar turn |
| image6 | Figure 5, size bands | Post 2.4, social card 3.4, sales one-pagers |
| image7 | Figure 6, definition gap | Carousel 3.1 spine, webinar, press kit |
| image8 | Figure 7, standard | One-pager 1.1, post 2.5, discovery guide, webinar close |
| image9 | Figure 8, limits | Post 3.7, method post, repo |

One production note: the figures are landscape at roughly 3:1. LinkedIn carousels and mobile feeds want square or portrait. Recutting from the SVGs is a real task, maybe an hour per format, and it is exactly the "channel-specific optimization" step the literature says teams skip.

---

## 6. Phase one

Thirty assets exceeds one production cycle. Build order below is set by dependency and by reach per hour of production, not by tier number.

**Produce first.**

1. **The one-page Change Evidence Standard (1.1).** The brief's back cover promises this by name and prints a URL for it. Until it exists, the flagship ships with a broken promise. A few hours of work using image8 as the base.
2. **The reproducibility package (1.5).** Near-complete. It has to be live before the brief reaches trade contacts, because the first question a skeptical reader asks is how the numbers were produced.
3. **The lead article (2.1).** The series opener. Nothing else in Tier 2 can be scheduled until the series has an established voice and cadence.
4. **The definition-gap carousel (3.1).** Highest reach per hour in the set. The outline is written in section 4.2; remaining work is recutting image7 to portrait.
5. **The webinar run of show and opening slides (4.1).** Promotion starts in week 2 of the launch sequence, so the shape of the session has to be settled well before the week 4 delivery date.

**Then the rest of section 3, in launch-sequence order.** The full tables stand as the production backlog. Nothing in Tiers 2 through 6 requires new research; every remaining asset is a reshaping of material the anchor already contains.

**Held for later.** The project self-audit worksheet (1.2) and the change classification codebook (1.4) are the two most directly usable assets in the program, and both deserve more design attention than a launch window allows. They are better as month-two releases that re-activate the campaign than as rushed launch-day files.

## 7. Launch sequence

Structured as a sequence rather than a publication date. Adapted for a solo publisher.

**Weeks minus 2 to 0, pre-launch**
- Finish the one-pager so the CTA on the brief's back cover resolves on day one.
- Publish the repo.
- Send the paper under embargo to two or three trade contacts and one or two practitioners who might quote it.

**Week 1 (launch)**
- Publish the paper and the announcement card (image1).
- Post the $10.6M hook (3.3) on day one. It is the best cold open.
- Follow with the carousel (3.1) on day three.

**Weeks 2 to 6 (sustain)**
- One article per week from Tier 2, each with its matching social card.
- Webinar in week 4, promoted from week 2.
- Trade pitch follow-ups.

**Weeks 7 to 12 (convert and extend)**
- Webinar cut-downs and on-demand replay.
- Role-specific one-pagers.
- Guest podcast appearances.

**Later (refresh)**
- The R in PADR. Add a second agency to the panel, or rerun with a trailing year of data, and the brief becomes an annual property rather than a one-off. Recurrence is what turns a single asset into a franchise, and it is the cheapest authority available once the method is already built.

---

## 8. Measurement

Match metrics to what each tier is for. Vanity numbers are a trap the literature names repeatedly.

- **Awareness (Tiers 3, 6):** citations and pickups, saves and shares rather than likes, unsolicited mentions from outside the outreach list.
- **Engagement (Tiers 2, 4):** completion depth on articles, webinar attendance quality rather than registration count, replay views.
- **Capture (Tier 1):** one-pager and worksheet downloads, and specifically who downloads them.
- **Downstream (Tier 5):** whether sales conversations reference the research, and whether the standard shows up in customer workflows.
- **Durability:** citations at six and twelve months. Original research compounds; this is the metric that separates it from campaign content.

---

## 9. Scaling the method

This program was produced from public records by one person, with no proprietary dataset and no subject-matter interviews, and a single ten-page brief still supports thirty mapped assets.

Two inputs would raise the ceiling rather than the count.

**Proprietary operational data.** Public records show how a transaction was filed. Platform and customer data show what practitioners actually do. Research built on the second cannot be reproduced by anyone working from the first, which is what makes an anchor durable instead of merely current.

**Subject-matter access.** Interviews with contractors, claims professionals, and underwriters would convert several Tier 2 and Tier 4 assets from analysis into reporting. Reporting carries a different kind of authority, and it produces quotable sources that extend the reach of every derivative built on top of it.

The atomization ratio holds either way. What changes is how defensible the anchor is, and therefore how long the derivatives keep earning.