startthat.work
startthat.work is a catalogue of analytical techniques and a process for choosing and running them. It takes a request and some material, chooses the analytical techniques that can run on that material, runs them in order, checks their findings and combines them into one answer, with the conclusions grouped by how well each is supported.
The name is startthat.work, which is also the address of this site. The skill that runs the process is named analysis.
The catalogue holds 236 techniques in 12 families. This site documents the process and the catalogue. It is generated from the design document, the schema and the catalogue file in the repository, so it changes only when those change.
Process
How a request and some material become one answer, in eight steps from intake to the answer.
Techniques
The 236 analytical techniques the process chooses from, with when each fits and what it needs from your material.
Dictionary
What each tag on a technique page means, and the checks and relations the process uses.
Source research
What techniques exist, what evidence supports them, how widely they are used and what the catalogue does not cover.
Limits
What the process does not do, what has not been tested and where its answers are weaker than they look.
Simulator
Replay one stored run step by step, to see what the process does and the answer it gives.
Skill
The skill that runs the process. Download it and install it in Claude or GitHub Copilot.
Use by hand
Follow the whole process yourself, without the skill, from the brief to the answer.
Status
Build and test updates.
Change log
What changed in each version of the design, and how to check the version of your copy of the skill.
Project record
Why the design is as it is: dated decisions, test evidence and the files.
Process at a glance
The process turns a request and some material into one answer. The user is asked only at the decision points. Each step links to the page that explains it.
| Step | What happens |
|---|---|
| 1. Intake | The user's input becomes a brief. Gaps become questions to the user. |
| 2. Rule out techniques | Each technique is tested against the brief. Techniques that cannot run on the material are removed. The rest are eligible. |
| 3. Choose analysis techniques | The model reads a list of the eligible techniques and chooses at most 10, with a reason for each. |
| 4. Plan analysis sequence | The chosen techniques are placed in four named groups in running order, each technique with a description of what it does and a note of whether it needs an additional source. Preparation techniques and checks are added. |
| 5. Run the techniques | 5.1 Examine the question. 5.2 Prepare the material, including any additional sources. 5.3 Analyse the material. |
| 6. Check the findings | Verification and challenge techniques run on the findings. The library checks are applied to each result. |
| 7. Combine the findings | The findings become conclusions. Each conclusion gets a confidence label and a one-line reason. |
| 8. Answer | The answer is given with a list of what it is based on. The user can ask for a follow-up round. |
Techniques at a glance
The 236 techniques are in 12 families. A family groups techniques by what they do for a request. The techniques page explains how to read a technique and lists them all.
| Family | Techniques | What the family is for |
|---|---|---|
| Frame the question | 6 | Turn a vague topic into a clear question, before any analysis starts. |
| Decompose and map | 21 | Break a topic into parts and show how the parts relate. |
| Compare | 17 | Set cases, periods or groups side by side to see what differs. |
| Trace cause and test claims | 34 | Look for causes, and test whether a claim holds. |
| Handle evidence | 29 | Judge how far to trust sources, and prepare them for use. |
| Project forward | 25 | Estimate what may happen next. |
| Assess actors, options and risks | 19 | Work out who is affected, what the choices are and what could go wrong. |
| Challenge | 15 | Test a conclusion by arguing against it. |
| Prepare and extract | 14 | Get the material ready, so that counts and themes describe what you think they do. |
| Measure and model | 16 | Analyse records with statistics and models. |
| Integrate across sources | 11 | Combine sources that overlap or disagree. |
| Check extracted results | 29 | Check results that came from coding or extraction, including model output. |
| Total | 236 |
The status page lists build and test updates.