# Weekly improvement loop.

Metagon Growth Starter Kit · file 6 of 7

Growth rarely comes from one big idea. It comes from small changes, made on purpose, written down and checked.

One change a week is enough. Over a year, that is about fifty changes you can explain.

---

## The routine.

Same day, same time, every week. 30 to 60 minutes.

1. **Look (10 minutes).** Open last week's numbers: the one result that matters (sales, bookings or requests), what
   you spent, and anything that changed. Compare with the week before.
2. **Check last week's change (10 minutes).** Did it do what you expected? Write the result in the log, even if the
   result is "too early to tell".
3. **Pick one change (15 minutes).** From the idea list, pick the change most likely to matter that you can finish
   this week. One. Not three.
4. **Write the reason (5 minutes).** Why this change, what you expect to happen, how you'll know, and when you'll check.
5. **Approve and do it.** If it costs money, the person who owns the budget approves it in writing first. Then make
   the change.

---

## Five rules that keep it honest.

- **One change at a time in each area,** so you know what caused what.
- **Give it time.** Wait for enough visitors and enough results before you judge. Small numbers lie.
- **Write the expectation first.** Before you see the result, not after.
- **Write down why you stop things.** A year from now you will want to know.
- **Bad weeks count too.** Note what happened outside: holidays, weather, a competitor's sale, the site going down.

---

## Where the ideas come from.

- Your answers to the growth check prompts (file 1).
- Questions customers asked this week.
- Every unticked box on the landing page checklist (file 3).
- Ads that stopped working.
- Things someone on your team does by hand every week.

---

## Idea list.

| Idea | Area | Effort | Why it might matter |
| --- | --- | --- | --- |
| | offer · page · ads · counting · follow-up | small · medium · big | |
| | | | |
| | | | |
| | | | |

---

## Decision log.

| Week | The change | Why | What we expect | How we'll know | Approved by | Check on | Result | Keep, change or stop |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |

Example row, made up:

| Week | The change | Why | What we expect | How we'll know | Approved by | Check on | Result | Keep, change or stop |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Week 1 | New headline on the booking page, in customers' words | In the stranger test (prompt 1), people misread the old one | More bookings from the same number of visitors | Bookings divided by page visits, compared with the four weeks before | The owner | Week 3 | | |

---

## When a change doesn't work.

That's a result too. Write down what you learned, undo the change if it hurt, and pick the next idea. A log full of
honest "stop" rows is worth more than a log nobody keeps.

---

Metagon Growth Starter Kit · metagon.co/growth-program