Three engagements from the last two years. Each one opens with the problem as the client described it, because that part is usually the interesting part.
Tidewater has sold technical marine gear out of Portland since 1984. Their customers are working captains who know exactly what they need, and the old site made them prove it: a PDF catalogue, a search box that matched part numbers only, and a checkout that timed out on a dock with one bar of signal.
The catalogue was not the problem. Customers were visiting product pages and leaving to phone an order in, because the site could not answer the two questions they always had — will this fit, and is it in stock. Both answers existed. Neither was on the page.
| Checkout completion | 41% → 68% |
|---|---|
| Phone orders | −38% |
| Largest contentful paint | 4.1s → 1.2s |
Measured over the first two quarters after launch, against the same season the year before.
| Time | Patient | Clinic | State |
|---|---|---|---|
| 09:00 | R. Alvarez | Northgate | seen |
| 09:30 | T. Nakamura | Northgate | in room |
| 10:15 | D. Okafor | Riverside | hold |
| 11:00 | M. Lindqvist | Northgate | confirmed |
Calder runs eleven clinics on a booking tool bought for a single site. It could not represent a blocked chair, the travel time between buildings, or the clinician who has to be in two places. Staff were keeping a parallel spreadsheet to work around it, and that spreadsheet was the most accurate thing in the building.
Watching a front-desk coordinator for two days made it obvious: they were not trying to book appointments, they were trying to avoid embarrassment. The rules that mattered were social and unwritten, and none of them were in the software.
| No-shows | −22% |
|---|---|
| Double-bookings | 0 |
| Parallel spreadsheets | Retired |
Across all eleven sites, two quarters after the first site went live. The other ten followed in three months.
Every legacy regional endpoint maps to this contract. Old keys keep working until you turn them off, and we will tell you which ones you can retire.
Ridgeline has bought eleven regional carriers in nine years. Each one described a shipment differently, and the only person who could answer a cross-regional question was a person in Portland. Growth had made the company dependent on one person's memory.
The data was not that bad. What was missing was an agreement about what a shipment is — eleven definitions, all locally reasonable, collectively unusable.
| Partner integration | 6 wks → 4 days |
|---|---|
| Systems behind one contract | 11 |
| Incidents traced to one owner | All of them |
The partner figure is the one Ridgeline quotes internally. New integrations now run without us in the room.
Pooled from the last thirty-six months, so the shape of the practice is visible rather than the best single result.
| Measure | Figure | What it counts |
|---|---|---|
| Launches shipped | 147 | Production releases, not prototypes or pitch demos. |
| Care clients who renewed | 91% | Across the last thirty-six months, measured at year one. |
| Median time to first release | 31 days | Kickoff to something a real user can open. |
| Work still running on our code | 132 | Live and maintained today. |
Halden is a fictional company and these are placeholder figures.
If one of these problems sounds like yours, say so and we will tell you honestly whether we are the right people.
Start a project