Leading design through Moladin's pivot

Context

Moladin is an Indonesian used-car marketplace. The part I joined builds the tools that make the marketplace move: agent-based car trading, affiliate-based consumer loans, and consumer loans. Later the company bet bigger on dealer financing — loan origination, loan management, collections, purging and inspection services. That pivot is the spine of this story.

I joined as Associate Design Lead, Ecosystem, owning design for the tools agents and affiliates use. The design organisation was about 20 people; I led 5 of them and reported to the then design lead. From 2023 I led design for the dealer financing pivot, and in September 2024 I became Design Lead for the group design team.

Role and team

The work spans two roles at the same company. As Associate Design Lead I led 5 designers and 3 copywriters in the Dealer Financing and Platform Tribe, owning delivery strategy, design review and stakeholder alignment. As Design Lead I now lead a 4-person team across 4 business verticals.

The number that matters most goes the wrong way, and it has dates: 15 designers in 2023, 10 by the end of 2023, 4 by August 2024. Design kept delivering with a quarter of the people.

The problem

This was not a failure you could see on a screen. A gap had opened between designers, product managers and business operations. Designers rarely saw how decisions were made, worked in silos on their own slice of a flow, and were unaware of the business dynamics or the challenges other teams faced.

It showed up twice: design questions arrived after the commercial shape of a feature had been decided without them, and reviews became arguments about preference rather than about what the business needed, because nobody had shared context to argue from.

What made this fixable was noticing who the user was. For these internal tools, the users are the ops and agent teams — the same people designers were negotiating with.

Constraints

• Headcount only went down. Every change had to work at four people, not fifteen.
• The scope did not shrink with the team. Dealer financing, Tuwaga, the Moladin web transformation and the Moladin Finance SME platform all had to keep moving.
• Nothing could be paused to reorganise. There was no quiet quarter to rebuild ways of working in; it had to happen inside delivery.
• Limited publication. Most of what this team built is internal or commercially sensitive, which is why this is a written case study — no screenshots are publishable.

What I did

I worked with several PMs to bring designers closer to the business, so they could see the field and empathise with the teams on the other side of the request — teams that were also the tools’ users.

Designers stopped being handed a specification and started being in the room where it came from: dealer and ops conversations, and the commercial reasoning behind a feature. The first question in a review moved from “does this look right” to “what is this supposed to change for the business, and what did we learn that supports it”.

The structural answer was how I distributed the four people. I chose a central pool that moves fluidly rather than one designer embedded per vertical — the argument for that is below. Three things made it work:

• Context sharing. Knowledge could not live in the people who had left. Decisions were written down and shared, which is what makes a designer movable without losing the thread.
• A maintainable design library. A library needing a dedicated maintainer is a liability at four people.
• AI for exploration and prototyping. To widen the option space early and reach something testable faster.

Key decisions and trade-offs

I did not embed one designer per vertical. Four designers and four verticals looks tidy, and it reads best in a review deck. I chose a central pool instead, because a pivot reallocates attention constantly: the vertical that matters most this quarter is not the one that mattered last quarter. A pool lets me move capacity to where the pressure is, with limited friction, and lets one designer cover for another without a handover. The cost is real — nobody becomes the undisputed expert for one vertical, and designers carry more context. I accepted shallower specialisation for a team that could absorb a departure.

I did not fight for headcount. Arguing for more designers was the obvious move and the least useful one. The company was pivoting; the constraint was real. I fitted our ways of working to the team we had, and was honest with stakeholders about what four people could carry.

I did not build a dedicated research function. At four people, a research team was never going to be resourced. Instead, business context was distributed to every designer, so judgement about the business lived across the team rather than in one specialist. The trade-off: no single owner of research rigour, in exchange for a pool where any designer can pick up any vertical’s work.

I did not let the design library expand to cover everything. A maintainable library means saying no to components. Visual consistency lost ground to keeping the library alive — and a lean library is what lets a designer move between verticals without relearning components.

I did not solve the designer–PM tension with a process document. A heavier process mid-pivot would have added coordination cost without context. The fix had to be proximity, not paperwork.

Outcome

The gap closed. Designers and PMs stopped working in silos and stopped blindly questioning each other, and began working together — with shared visibility into what the business faced.

Against the shrinking team, the portfolio still shipped:

• Dealer financing went from 0 to 5,000+ users.
• Tuwaga went from 0 to 1,000+ potential leads.
• The Moladin web transformation moved revenue from hundreds of thousands to hundreds of millions of rupiah per month.
• The Moladin Finance SME platform manages hundreds of leads and loan submissions, with a companion agent app generating 100+ leads per month.

I would not claim a causal line from the ways-of-working change to any single number. What I would claim is that the same team kept delivering four initiatives while it lost most of its people, and that the silo problem stopped being what slowed decisions down.

What I’d do differently

I framed the silo problem as a designer problem for too long. It was a company problem that happened to be visible in design. Designers did not know how decisions were made because the decision-making was not legible to anyone outside the room — so I translated, and translating for a while made the team comfortable rather than self-sufficient. I should have pushed business context out earlier, especially into finance and ops, whose constraints reached design last, as vetoes.

I measured the change through the quality of arguments in reviews: a real signal, but a private one. I should have left something observable behind — a record of what each vertical’s tools were meant to change and what was measured afterwards, making the team’s judgement legible to whoever inherited it.