All work

Case 03 · ERIKS · 3 min

Case 03 · Industrial B2B

The part is cheap.
The stopped line is not.

Research and vision concepts for ERIKS Gateway, a self-service portal for the people who keep factories running. A production line that stops can cost more in an hour than a year of the parts that keep it turning.

Field research Job stories Service design Interaction design

Role  UX designer — research, then design Client  ERIKS — built with the ZAMRO team With  Mirabeau Duration  Several months Deliverable  A vision document

Two people in blue ERIKS shirts at an assembly cell. One leans in with a hand on a part under the tooling. The other stands behind him, watching the same point.
Sometimes the cheapest part is the reason nothing else is moving.

Where this started

You cannot specify a job you have never watched.

None of it had ever been properly validated. There were opinions about what these customers needed, and nothing to check them against. So before I designed anything, appointments were made at customers’ plants, and a colleague and I went to look.

We interviewed on site and came back with the job as it is actually done: the steps people take, the tools they already trust, how a decision gets made and who has to be in it. Then, role by role, what each person is trying to achieve.

The first finding was not that a plant holds several roles, but how differently each one meets the same stopped line. The engineer at the machine, the buyer in advance, the manager across plants. Tell all three the same thing and you are wrong twice.

A line drawing of a factory in cross-section with a sawtooth roof and a smoking chimney. Three windows show three people: on the shop floor a worker in a blue hard hat and overalls holding a clipboard among the machines, high in the tower a figure in a tie pointing at a rising chart, and below them someone on a headset at a desk.

The same thing, in their words

Situations, not features.

A feature list invites argument about the feature; a job story invites argument about the situation. A couple of example job stories below, out of a much longer list.

Engineer

When a part on my line has failed, I want the order for that one part in front of me, where it is and when it lands, so I know whether I am opening the machine today or planning around it.

Buyer & planner

When I am planning ahead, I want to see which orders have slipped before anybody asks me, so I can move a delivery or a maintenance window while there is still time to move it.

Procurement manager

When delivery performance drops sharply at one site, I want to hear about it without logging in, so a bad month does not reach me once it has become a bad quarter.

The second finding

They did not need another portal.

Research earns its keep when it comes back with something the brief did not ask for. That is where the budget stops going into features nobody would have used unless they had to.

Most industrial-size plants already run portals for everything, each with its own login and its own password to lose. One more would have been another account rather than less work.

So I changed the question: not what this one should do, but how much of it needed to be a portal at all. Part of the answer was email — a different one per role, carrying the few things that role can act on without going anywhere.

The portal that was left starts with what is going wrong. If uptime is what is at stake, the orders worth showing are the ones threatening it, so the overview leads with those and the full list sits one level down. Every page was held to the same rule: manage by exception.

A line drawing in two stacked scenes. Above, a man stands in a corridor holding a heavy ring of keys, facing six identical closed doors that recede into the distance, each with its own padlock and a blank sign plate. Below, the same man sits at his own desk with a laptop and a mug while a blue envelope sails in through an open window behind him.

What management got to react to

A clickable prototype, built to be argued with.

What we handed over at the end of the research was a vision document and the presentation around it. These fifteen screens are that vision made clickable: something management could move through and react to rather than read, showing what we believed an answer to their customers’ situations could look like — and what ERIKS’ own expertise and advantage could be worth inside it. It stands for a direction, not a build.

The first frame of the prototype: a mock desktop mail client with the monthly ERIKS update open in the reading pane.
The first screen is a mail client, not a login. The monthly update arrives where the work already happens.
The dashboard: projects grouped under Delayed, On track and Done with coloured status chips, and a panel showing an in-contract versus out-of-contract split.
The overview opens on what is late. Two delayed projects, above everything on track.
A milestone table: two completed rows with dates, a current row with an active Notify owner button, a future row with a greyed Complete button, and a last row with no due date at all.
A project as it really stands. Two done, one waiting, one not started.

Next

The other cases.

Three years of paperless maintenance in a KLM hangar, and a self-service portal that had to read in both directions.

Back to all work