Every post, against your own median.
Harbor measures each post at five ages and compares it with the brand’s median on the same platform at the same age. The weekly digest turns that into what to do next, citing the posts.
Five checkpoints, one fair comparison.
A post’s views at 24 hours mean little on their own. Harbor snapshots every post at 1h, 6h, 24h, 3 days and 7 days, and draws its curve against the median of the brand’s other posts on that platform at the same age.
- Each brand and platform keeps its own baseline
- Every snapshot is kept, metric by metric
- Imported posts join in at their true age
A verdict you can act on.
At 1.3× the median or more, a post outperformed. Under 0.7×, it underperformed. In between, it met the brand’s usual. With fewer than 3 comparable posts, Harbor says inconclusive instead of guessing.
- A first outperform can start an automation
- Verdicts feed the voice refresh and the digest
- LinkedIn shows metrics as unavailable until partner access, never estimated
- 3 weeks, 2 rewrites, 1 feature.Waiting for 24h
- Most founders with several products hit the same wall…Waiting for 24h
- Pricing is per brand, not per seat. Here’s why.Waiting for 24h
- What we look for in a design partnerWaiting for 24h
- New changelog is up: harbor.example/changelogWaiting for 24h
- This brand’s third post on ThreadsWaiting for 24h
A digest that says what to do next.
Once a week, or whenever you ask, Harbor reads every post’s curve and writes what worked, what didn’t, the patterns, and what to do next. Each finding carries a confidence and cites the posts behind it. The weekly plan reads it before suggesting anything.
- Run it by hand or on a schedule
- Every finding links to its posts
- The weekly plan reuses a digest under 24 hours old
Best times, from your own posts.
A 7 × 24 grid per brand, in one hue: darker is better. Scores come from the brand’s own posts, blended with platform norms until there’s enough history. The top 5 are outlined, and each cell shows its score and what it’s based on.
- Every brand has its own map
- The same scores shade the calendar
- Agents use them to propose slots
- Sun 7p100
- Thu 9a98
- Mon 9a96
- Sun 8p96
- Wed 9a92
What people keep saying.
Harbor reads replies 24 hours and 3 days after each post and groups them by kind (questions, objections, praise and requests) and by what they have in common. Each theme becomes a signal that ideas and the weekly plan can cite, and any theme can be saved to Brand Memory.
- Instagram and Threads replies, plus X when paid reads are on
- LinkedIn’s API doesn’t allow reading replies yet
- Each theme counts its replies and the posts they came from
Plan next week in an afternoon.
Self-host Harbor free, or get early access to Harbor Cloud. Either way, nothing posts until you say so.