Tech direction · October 2026

One product,
owned close to home

Where we put our effort, how we build, and what we need to get there.

Alexander · Tech Lead

Who we are

We are a tech company in scale-up

Our customers are coach operators.
Our product is technology.

What we own The platform, the API and the data What we build is the company's core asset.
What it means Tech is the product, not overhead Investing in technology is investing in what we sell.

01

Where we stand

Where we stand

What customers are telling us

#1Performance is the top reason customers are unhappy
35%of shipped features are in use across ten key customers
−20NPS excluding Tide: a product problem we can fix
V5in UAT, production targeted for mid to late October
Development is moving to local ownership in Oslo. The rest of this talk is about how we make that work.

Sources: CSAT March 2026, feature adoption across ten key customers, roadmap board 6 October 2026.

The principle

One product.
A small team that owns it.

Five to ten developers in Oslo can build and maintain Ferdia. That works because of how we build, not just who builds.

02

How we build

A cleaner foundation

Each layer has one job

  • Change one part without breaking another. Clear lines between data, logic and screens.
  • Web and CoDriver move at their own pace. Each gets a backend shaped for its needs.
  • Faster screens. Each backend returns what the screen needs, nothing more.
WebApp shell and apps
CoDriverDriver app
Partners and AI agentsIntegrations, MCP
BFFBackend for frontend
BFABackend for app
APIOne API, internal and public
Database

The app shell

Load what you use, nothing more

  • Shared things live in one shell: navigation, login, settings and notifications.
  • Apps load only when needed. A planner opening Planning doesn't download Sales.
  • Mirrors how we sell: modules on one platform, each owned and shipped on its own.
Ferdia Navigation · Login · Settings · Notifications
PlanningOpen now
SalesLoads when opened
CoPlannerLoads when opened
Customer PortalLoads when opened
NetworkLoads when opened
InsightLoads when opened

Start small

No rewrite. Four steps, each useful on its own.

  1. V5 shipsInto production, as planned.
  2. Separate repositoriesMonorepo, repos with clear responsibilities.
  3. App shellWraps what already exists.
  4. Apps move inOne at a time, when ready.
Every step delivers value before the next one starts. We can pause between steps without leaving anything half-built.

One API

One API, as a product

Today we keep an internal API and a public API. Tomorrow there is one, and our own product uses it too.

  • One surfaceInternal and public mergedLess to build, test and maintain.
  • We use what we sellOur product runs on itWe find the gaps before customers do.
  • TiersSomething new to sellMore data, more features, no rate limits.
  • AI includedMCP on the same APIEvery feature is available to AI agents, without extra work.

Seeing what happens

One record of what happens in Ferdia

Activity stream Trips, orders, invoices and changes, as they happen Built on Activity Streams 2.0, an open W3C standard.
SupportA trail of what happened to a trip or an order.
ProductAdoption at 30 and 90 days, measured automatically.
IntegratorsWebhooks later, so partners stop polling.
EveryoneA company-wide dashboard with the same numbers.

The golden path

One paved way to build, test and ship

  • Fewer decisionsSetup, testing and deployment are solved once, not in every project.
  • Fewer regressionsEvery app passes the same checks before it reaches customers.
  • Faster onboardingNew developers are productive in days, not months.
This is what makes a small team sustainable. Five to ten people can own one product when the way we work is shared, written down and automated.

03

The road there

The transition

A gradual handover, invisible to customers

  1. PairOslo and Dhaka build the app together.
  2. OwnOslo leads, Dhaka supports.
  3. Hand overOslo owns it, Dhaka steps back.
  • App by app, repeated until we are fully across.
  • In parallel with existing development, with no gap in capacity or knowledge.
  • Committed roadmap items stay on track. The transition is planned around them, not instead of them.

What we gain

Shorter time from request to production

  1. RequestHubSpot
  2. RoadmapJira ROAD
  3. SprintJira PROD
  4. ReleaseFortnightly
  5. Adoption30 and 90 days
Now We measure today's lead time Before anything changes, so the comparison is honest.
Then The dashboard shows it improving Visible to everyone, not hidden in a report.

What changes for you

What this means in your day

  • SalesThe roadmap holds, and API tiers give you something new to sell.
  • MarketingA real technology story against the AI claims of new challengers.
  • SupportA trail of what happened, and fewer regressions to chase.
  • ProductDevelopers nearby, and a shorter loop from request to release.
  • LeadershipA shared dashboard and clear milestones for the transition.
  • CustomersFaster screens, and no disruption while we change.

Design

Design is part of the product

Today we have one UX consultant at 20%, working mostly on UI. Yet much of what customers tell us is about design:

  • Pricing and order UX
  • Visual Planner friction
  • 35% feature adoption
  • 3 churned before go-live
Start 50% UX and 50% UI Covers the most urgent product and interface work.
Goal UX, UI and service design Three roles. Service design covers onboarding, where we lose customers before go-live.

What I need from you

Three things

  1. UnderstandingWe are a tech company. The product is what we build.
  2. PatienceThe transition takes time, and results come gradually.
  3. A careful but committed budgetFor local development, local ownership and design.

Thank you

Questions

Alexander · Tech Lead

Ferdia - One product, owned close to home

By Alexander Vassbotn Røyne-Helgesen

Ferdia - One product, owned close to home

  • 86

More from Alexander Vassbotn Røyne-Helgesen