- senior engineers
- 60
- strategic decisions shipped
- 4
- NPS vs office strategy week
- +24
- continue in the cohort
- 8
This case is an anonymised composite of three engagements with large fintech companies that approached us in 2023 and 2024. "Stripe Mediterranean" is the internal working code under which we ran the project at the planning stage. The real client is a global payments fintech at Series E with offices in seven countries. Financial figures and personally identifying details have been removed; the metrics — NPS, ship velocity, retention — are original and aggregated. Under NDA we can show the original documents and contract on request.
Context
The client
A payments fintech, Series E, $2.6bn ARR, 4,200 employees. The engineering org is 1,200 people split into 14 product engineering groups. The senior engineering layer — staff and principal engineers — is 80 people, gathered for an annual strategy week. Until 2023 the format was standard: four days in San Francisco, two in London, conference rooms with whiteboards and pizza. The central question of the strategy week is the roadmap for the next fiscal year and the arbitration between groups over how to split engineering capacity.
The pain
The VP Engineering came with a specific pain: "by day four in the conference room we make roadmap decisions not because they are right but because people want to go home." That phrase came up verbatim on the first discovery call in April 2024. After the 2023 strategy week the company ran a retrospective survey — the three most frequent answers were that the format had exhausted them by Wednesday, that priority decisions were made because everyone was tired of arguing, and that the Dublin and Singapore teams flew home feeling under-represented. NPS sat at 38, critically low for a senior engineering corps.
Why us
Discovery took five weeks. The VP Engineering reached us through an executive coach we had advised on another programme. The main constraint he framed was that the format had to hold for four days without an energy drop and still allow hard prioritisation calls. We proposed Palma de Mallorca as the base: October, peak-mild weather, short passages between bays, no stressful winds. Four days — an arrivals and brief day, two days of strategy work on the water with a facilitator, and a closing day for prioritisation and signing off the roadmap decisions. The contract was signed six weeks later.
What we did
Stage 1
Brief and route
The route ran along the south coast of Mallorca — four Bali 4.4 catamarans, 15 people each. We started from the Palma marina, with passages to Cala d'Or and Porto Cristo and overnights in sheltered bays. A deep discovery with the VP Engineering and three engineering leads let us understand the structure of the priority conflicts, so we could place the right people in the right groups on each yacht. Of the 60 people, 23 were staff engineers and 37 principal engineers, plus 12 invited engineering managers. We grouped by cross-product mixing to break the vertical silos.
Stage 2
Logistics
Pre- and post-event accommodation was the Castillo Hotel Son Vida in Palma (two nights before departure, one after). Transfers from Palma airport ran as three minibuses in a 12:00 to 17:00 window on arrival day. Each yacht carried a Starlink Mini for a stable connection — a requirement from on-call engineers who cannot afford to drop coverage. Catering was lunches aboard and dinners in Cala d'Or (Bistro 1925 and Restaurante Vista Mar) and Porto Cristo (Mhares Sea Club). Special-diet requests ran higher than usual — Bay Area engineering favours veganism and intermittent fasting — so we added a third provisioning option on each yacht.
Stage 3
Programme afloat and ashore
The programme was built around three three-hour strategy blocks on the water. Block 1 was "What we won't build in 2025", run as a structured debate between two groups on two adjacent yachts using walkie-talkies. Block 2 was a cross-team dependency map, worked in mixed groups around a physical whiteboard on deck with the result photographed on an iPad. Block 3 was the finals priority list, on the closing day, at anchor in Cala d'Or. Between the blocks were three hours of cruising with no agenda. That turned out to matter — the unfacilitated pause gave space for the informal conversations that happen at office coffee points. We also added a pre-event 90-minute sail induction in the marina so nobody felt nervous on the first sail.
Stage 4
Results and KPIs
NPS 14 days later was 62 (against 38 for the previous San Francisco strategy week — a lift of 24). Four strategic roadmap decisions were signed off right on the final yacht; all four reached production within 90 days, a historic record for this company. Of the 60 senior engineers, 8 became part of a peer cohort that meets monthly over Zoom for architecture review. The VP Engineering recorded a retrospective post: "the difference is that we came back with decisions made, not with tabled discussions." Inside the company the programme has already become an informal reference format for other departments — sales, customer success — which now want a similar strategy week in 2025.
“We were used to running our strategy week in San Francisco, in a glass meeting room. Effective, yes — but by the end of the fourth day people are wrung out and make decisions because they want to go home. On the yacht it was the opposite: by day four there was still energy for three additional roadmap decisions.”
Photos from the programme
Lessons
- A stable connection is critical for tech teams. A Starlink Mini on the mast is not overkill — it is a precondition. Engineers who cannot answer a page cannot take part in strategy.
- Cruising windows between work blocks do not waste time, they add value. In an office strategy week a "pause" is a coffee point. On a yacht a pause is two hours of informal conversation among a group of four or five people who rarely cross paths. Those conversations turned out to source most of the unplanned roadmap insights.
- No more than three work blocks a day, and no more than three hours each. A fourth block lowers decision quality regardless of the format. That law holds both in glass meeting rooms and on the water.
- The final priority list is best signed at anchor in a bay rather than at a hotel. The physical isolation — two hours without a connection — is a powerful signal to the mind that the decision is final.
Facilitation and working with hidden conflict
The senior engineering layer of a fintech is a group whose accumulated technical disagreements have gone unspoken for years. At the previous San Francisco strategy week several of those disagreements surfaced as open confrontation on the Wednesday, and two engineering managers left the next day. We built the programme on the assumption that such conflicts exist and will surface — and deliberately invited two certified executive coaches with experience facilitating tech teams to join in a shadow role for the first two days. They did not run sessions directly, but they helped the VP Engineering and the facilitator respond well to escalations in the moment.
Of the six escalations recorded over four days, all six were resolved before the conclusion block — one of our key operational success metrics. The escalations arose around platform versus application team priorities, the split of on-call load between regions, and contested architectural decisions from the previous year. One of the coaches put it this way: "the format on the water sped up resolution because people had nowhere to walk away from the conversation — you cannot leave the meeting room and pretend the conflict never happened."
Financial breakdown
The budget for 60 people over 4 days was €138,000 excluding flights. Per participant that was €2,300. The split: charter — 38 %; Castillo Hotel accommodation — 19 %; catering — 14 %; programme (facilitator plus additional facilitation support) — 12 %; Starlink and tech setup on the yachts — 4 %; logistics — 5 %; insurance and fees — 4 %; contingency — 4 %. Compared with the San Francisco strategy week, where the pure conference cost was €1,600 per participant with no learning ROI, this is a comparable investment with a fundamentally different outcome.
Ninety days later
All four signed roadmap decisions reached production. One of them — a move to a new observability platform — saved the company an internally estimated €1.4m a year in operational cost. That single project paid back the whole cost of the offsite tenfold in the first quarter after rollout. It is not an argument for a sales proposal, but it is the argument the VP Engineering used to defend the 2025 strategy week budget to the CFO. Finance approved a repeat of the format with a 20 % larger budget "for experiments with an expanded programme".
An additional metric we recorded six months on: retention in the senior engineering layer rose by 11 points against the previous cohort. That is a statistically meaningful change, because baseline senior-engineer retention in fintech of this kind swings within 4 to 6 % a year. The whole effect cannot be attributed to the trip, but the recruiting team noted that in exit surveys the mention of "lack of connection with the team" fell threefold against the previous year.
