Colorkrew · Workflows

From a one-line brief to a product specification

Colorkrew wanted an approvals product of its own. I was given a basic concept. What the product did, how it worked and what went into the first release were mine to define.

Role
Lead Product Designer
Dates
2023 to 2024
Product
Colorkrew Workflows, a cloud form builder with approval tracking and auditing
Team
Sole designer, a product owner and around three engineers
Scope
Specification, design and delivery support
Status
Released internally for testing. Now a live product built on the structure and flows I specified
From a paper form passed desk to desk, to a request with named approvers and a record of every step. Recreated for this case study; details changed.

The problem

Colorkrew makes workplace software for businesses. Like many companies, its own admin team still ran approvals on stamped paper forms passed from desk to desk.

The main digital alternative was hard to use, according to our own admin staff. Colorkrew wanted a product of its own, used internally first and offered to customers later.

The brief was a basic concept. Everything else had to be defined.

Constraints

  • One designer, many products. Colorkrew had around 10 products in development. For part of this time I was the only designer across them.
  • A small team. A product owner and around three engineers.
  • Part of a suite. Workflows had to fit Colorkrew's shared account system and sit alongside the other products.

My role

I wrote the specification and designed the product. I prioritised features with the product owner, weighing effort against value, and directed the engineers building the schemas and prototypes. The product owner set business priorities; I set feature priorities within them.


Decision 1: start linear

Approval processes get complicated quickly: if the amount is over a limit, go to finance; if it's from a certain team, skip a step. I kept the first version strictly linear. A request goes to approver A, then B, then C. No conditional branches.

The form fields were simple too, and connections to Colorkrew's other products were left out of the first release.

The linear flow we built, next to the conditional flow we chose not to build yet.

Decision 2: ready-made flows for common requests

Most requests are routine, and time off is the obvious one. I designed ready-made approval flows for common cases so teams could start without building anything.

A ready-made time off flow. Recreated for this case study; details changed.

Decision 3: a form builder admin staff could use alone

Admin staff needed to create and change forms without asking an engineer. I designed a drag-and-drop form builder. Some of the drag-and-drop interactions were simplified for the first release to keep it buildable with a small team.

The form builder: pick a field type, drop it into the form. Recreated for this case study; details changed.
The form builder in the live product, from Colorkrew's promotional video. Its visual design has changed since I left.

Holding the scope

The harder part was keeping the first version small. Feature requests came often, usually when someone saw a competitor's app with a feature we didn't have. Working through an effort and value grid with the product owner gave us a shared way to decide what went into the first release and what waited.

In the first release

  • Linear approvals
  • Prebuilt flows for common requests
  • The form builder, with simplified drag and drop
  • Simple field types

After validation

  • Conditional branches
  • Links to other Colorkrew products
  • Richer drag and drop

Outcome

  • A full specification, and a first version released internally for testing.
  • Workflows is now a live Colorkrew product. Its structure and flows are the ones I specified; the visual design has changed since I left.
Workflows today, on desktop and mobile, from Colorkrew's promotional video. The structure and flows follow my specification; the visual design is the current team's. Conditional branches, deferred from the first release, have since shipped.

What I'd do differently

Feedback from internal testing was mostly anecdotal, and new requests were often driven by competitors rather than by how people used the product. Next time I'd agree a few simple measures with the admin team before release, so decisions about what to build next rest on evidence.