Case StudySubscription Overview
Giving CS one subscription view, instead of nine tabs and a second CRM
Instafleet is Instacar's internal CRM, and Subscriptions is the Customer Success board. Answering a routine customer question — am I paid up, when does my contract end? — meant reconstructing the picture across nine tabs, then opening a second CRM that held the notes.
Overview
One Overview tab that answers the customer's question without navigating away.
The information existed. The product just made every agent assemble it themselves, on every call — and kept a second CRM alive to hold the parts it couldn't show.
This project wasn't assigned to me. There was no ticket and no brief. I noticed agents working around the tool, proposed the project, backed it with research across every team that touches a subscription, and designed it solo against the migration deadline.
The Problem
A customer calls to ask whether they're paid up and when their contract ends. To answer, an agent rebuilds the picture across nine tabs, then switches to a second CRM for the notes. One agent opened ten separate tickets to reconstruct a single conversation.
That cost showed up in several places at once: time on every enquiry, repeated per call and per agent; two licences and two training paths for staff; a retention lever — a free month offered near the end of month-to-month contracts — buried in a list of dates; and an operational risk, cars released to customers who owed money, because unpaid-balance holds weren't visible.
The clearest signal was in a single free-text box: fifteen unprompted requests for missing information. People don't write fifteen wishes about a tool they've made peace with.
Discovery
I surveyed 24 people across all five teams that touch a subscription — Customer Success, Finance, Sales, Management, and Credit — because I didn't trust my own hypothesis. Three things came back that I would have got wrong.
I assumed the record should lead with total contract value, the biggest number on the page. The team uses monthly value, because that's the language they use with customers.
What kept the old CRM alive wasn't a feature — it was notes. The free-text record of what was said was the most common reason anyone opened it.
A layout question surfaced a gap nobody had specified: three people in Credit and Finance independently asked whether a customer was in legal proceedings. The field didn't exist.
The Reframe
The data wasn't missing. The product made people assemble it themselves, every single time. So Overview became the first tab and the default landing view — a place to understand a subscription, not just a form to read.
Design Decisions
Identity first. Each subscription is titled by vehicle, customer, and plate, with status and type as badges — removing the orientation step at the start of every call.
Five metrics, chosen by evidence. The free nineteenth month got its own tile despite ranking seventh by frequency, because it's the retention lever on month-to-month contracts and both Sales and Finance named it directly.
Billing leads with what people use. Payment method and date sit above bank and IBAN — against my instinct, because reference data looks important but is read rarely.
Linked items replaced leaving the tool. Attached tickets and bookings now appear inline, removing one of the three reasons to open the second CRM.
Notes take the top of the activity column. The biggest gap got the best position.
Reflection
The design wasn't the hard part. The case was. Nobody asked for this, so it only existed if I could justify the engineering time against a roadmap that already had owners.
I argued it on user pain, and it worked — but that was the second-best argument. The money would have moved faster: the licence line, the agent hours, the renewal window. Cost and risk get a slot on a roadmap; friction gets sympathy.
The limitation: a survey measures what people say they look for, not what they do. The content order is a well-evidenced hypothesis, not a settled answer — the next step is an observation session with a few agents about four weeks after launch.
Outcome
- Surfaced a net-new risk requirement nobody had specified — it reached engineering before build, not as rework after.
- Delivered entirely within the existing design system, with no new components.
- Reordered the record around evidence of use, not assumptions about importance.
Quantified impact is being confirmed with Finance and CS: the licence cost removed, agent hours returned and costed, onboarding time for one tool instead of two, renewal contact rate inside the retention window, and vehicles released against an unpaid balance. The baseline for handling time is being captured before the screen ships.