Case study · 2023 · Lagerhus
An onboarding that admits the first week is the hard part
Rebuilding activation for a B2B analytics tool where value arrives only after someone connects a warehouse they do not administer.
Overview
Lagerhus sold warehouse analytics to operations teams who could not connect a warehouse without help from IT. Trials expired while tickets sat in a queue nobody on our side could see.
I led design on activation: the first session, the connection flow, and the week of waiting that neither sales nor support had ever treated as product.
Problem
The product was bought by an operations lead and installed by a systems administrator who had not been in the sales conversation and had no reason to prioritise it.
The brief was to shorten onboarding. The problem was that the person clicking through onboarding could not finish it alone, and the product had no way to say so.
Research
Nineteen trial accounts watched from signup to expiry, twelve interviews split between buyers and the administrators they delegated to, and a read of every support thread from a full quarter.
The finding that reframed the project: median time from signup to first connected warehouse was eleven days, and nine of those were spent waiting on a person who had never seen the product.
Process and explorations
We tried three shapes: a sample dataset that let buyers explore before connecting anything, a guided handoff that brought the administrator in as a named participant, and a stripped connection flow aimed at doing it yourself.
Sample data tested best in the first session and worst in week two — nothing carried over. We shipped the guided handoff, and kept the sample dataset only for the buyer's first ten minutes.
Final design
The first screen asks one question: do you administer this warehouse system. Both answers lead somewhere useful, and only one of them is onboarding.
Delegation became a first-class object. The buyer names an administrator, the product writes the request, and both sides watch the same status without having to ask each other.
Prototype
We ran the handoff as a coded prototype against six real trial accounts, with support staff playing themselves. Two of the four emails we had written turned out to be unnecessary.
Impact
Measured against the two quarters before launch, across self-serve trials.
- 11 → 3
- Median days to first connected warehouse
- +34%
- Trials reaching an active dashboard
- −58%
- Onboarding support tickets
Reflection
We spent two months designing a better empty state for a screen the buyer could not reach. The blocker was two roles and one missing message between them.
The thing I keep: activation metrics that stop at the account boundary will hide the person your product actually depends on.