Review a feature's results after launch and decide what comes next

Write a post-launch review of the feature below, for {{decision_maker}}, so they can decide whether to keep investing in it, change it or roll it back. It should read as an honest account of what happened against what we said would happen, not a celebration of the launch.

What the feature does and when it launched: {{feature}}

The goals and success metrics we set before launch, with targets:
"""
{{goals}}
"""

Results since launch (usage data and metric changes, with the date range):
"""
{{results}}
"""

Customer and team feedback (support tickets, survey comments, sales notes, interview notes):
"""
{{feedback}}
"""

What it cost to build and what it costs to run, if known: {{cost}}

Use these headings:
1. Summary: three sentences covering whether it met its goals, the main reason, and the decision needed.
2. Goals against results: a table with the columns Goal, Target, Actual, Met (yes, partly or no), Source. Show the arithmetic for any rate you calculate.
3. Who uses it and who does not: adoption by segment or plan, where the data allows it.
4. What customers say: up to 5 themes, each with a short quote and how many times it came up.
5. What we got wrong: assumptions behind the goals that the results contradict.
6. Options: continue as is, improve (with the specific change), expand, or roll back, each with the evidence for it and what it would cost.
7. Recommendation, labeled as the writer's view, with the one metric to watch next and when to review again.

Keep it under 800 words, not counting the table.

For example, the form of one table row: "Accounts using saved views weekly | 25% of active accounts by week 6 | 14% (812 of 5,800) | no | product analytics, weeks 1 to 6". This shows the form only; use the real numbers.

A good review makes the gap between plan and result impossible to miss, so report misses as plainly as wins. Use only the numbers and quotes above. Do not move the goalposts by swapping in a metric that looks better than the one we set; if you think a different metric matters more, say so separately under Options. Do not credit the feature with a change that other launches, campaigns or seasons in the same period could explain; name the other possible causes instead.

If a goal had no target, or the results do not cover a goal, mark it "not measured" in the table rather than estimating it. If the data period is too short to judge a goal, say so and give the date when it can be judged.