Extract a data model from one form

kind: drill

Here is a sign-in sheet as it hangs in a club:

Trainingsliste                       Gruppe: ____________
Datum: __________   Trainer/in: ______________________

Nr | Name                | Beitrag bezahlt | anwesend | Bemerkung
 1 |                     |  ja / nein      |    x     |
 2 |                     |  ja / nein      |    x     |
...
20 |                     |  ja / nein      |    x     |

Unterschrift Trainer/in: ______________________

Extract the candidate attributes with name, type, range and whether they are mandatory. Then write down three questions this sheet raises.

Solution

Attributes: gruppe (text, mandatory), datum (date, mandatory), trainer (text, mandatory), nummer (integer 1–20), name (text), beitrag_bezahlt (boolean), anwesend (boolean, recorded as a cross — absence is no cross, not a "no"), bemerkung (free text, optional), unterschrift (signature, mandatory).

Two structural findings come for free. The sheet caps a group at 20 — is that the paper or the rule? And anwesend has no third state, so "excused" currently lives in Bemerkung as free text.

Three questions worth taking to the interview:

  1. Is the maximum of 20 a property of groups or of the sheet?

  2. What does an empty anwesend mean — absent, or not yet recorded?

  3. What is written in Bemerkung, and does anybody read it afterwards?

Note that all three come with evidence attached. That is what makes them good questions.

Fix six interview questions

kind: drill

Mark the suggestive and the closed questions, and rewrite them:

  1. "You would like an app, wouldn’t you?"

  2. "Is the paper list annoying?"

  3. "How many members are in a group?"

  4. "Do you want the export as PDF or as Excel?"

  5. "Should the system be fast?"

  6. "Is it not a problem that the list is often incomplete?"

Solution

1, 2, 4 and 6 are suggestive — each offers its own answer, and 4 additionally assumes that an export is wanted at all. 3 is closed but legitimate: a number is exactly what is being asked for, and it is better taken from the documents anyway. 5 is unusable because "fast" has no unit.

Rewritten:

  1. "Walk me through how you record attendance today."

  2. "What is the most annoying part of the current process?"

  3. (keep — or read it off the sheet and confirm it)

  4. "What happens to the list after training?" — the export format follows from the answer.

  5. "How long may it take before it disturbs your work?"

  6. "How complete is the list at the end of a week, and what follows from a gap?"

The pattern: replace your hypothesis with a question about the last concrete occurrence.

Decide which technique fits

kind: drill

Name the technique and justify it in one sentence:

  1. The trainers say the paper list works fine, but three of them keep a private Excel sheet.

  2. You need to know how many of the 400 members would use online booking.

  3. Nobody can say which fields the club’s annual report needs.

  4. Two trainers describe the cancellation rule differently.

Solution

  1. Beobachtung — description and behaviour disagree, and only watching resolves that; an interview returns the description again.

  2. Fragebogen — a question about quantity across a population too large to interview, with closed answers and no need for follow-ups.

  3. Dokumentenanalyse — last year’s report exists; its fields are the answer, and reading it costs nobody anything.

  4. Interview, with both of them and the document in front of you — the contradiction is exactly what an interview is for, and the statute probably settles it.

Write the assumption list

kind: drill

From the exercises above, write five assumptions in the form We assume X. Source: Y. To be confirmed by: Z.

Solution

For example:

We assume a group has at most 20 members.
  Source: the sign-in sheet has 20 rows. To be confirmed by: the head trainer.

We assume an empty "anwesend" means absent.
  Source: assumption, no document says so. To be confirmed by: any trainer.

We assume "excused" is not recorded separately today.
  Source: no field on the sheet, only Bemerkung. To be confirmed by: interview.

Two properties make such a list worth writing: each line has a source, so a reader can judge how solid it is, and each names who can settle it, so it can actually be closed. A list without those two columns turns into a wish list nobody works through.

Run the document analysis for your own project

kind: project

Collect the documents your client already works with, catalogue them, extract the field list, build the first glossary, and bring the open questions to your first interview.