Case study
Every Customer Conversation, in One Place
The same three angry emails, a warehouse of every touchpoint since the company began, and a services flow designed to make the chase unnecessary.
Haim Repael · 2 min read

STAR AT A GLANCE
- Situation
- Customers who ordered services kept sending the same three emails, and we thought we knew why.
- Task
- Find where customers actually got stuck, from evidence rather than anecdotes.
- Action
- Unified every customer touchpoint since the company was founded into one warehouse, classified it, and turned the findings into the brief for a new ordering flow.
- Result
- A transparent services flow with automatic stage updates, validated in stakeholder sessions and interviews. In development; not yet measured.
What is happening with this? Where does my project stand? Why isn't anyone answering me?
Customers who ordered services from us kept sending the same three emails. We thought we knew why. We didn't.
SituationA familiar complaint with an unfamiliar cause
The story inside the company was simple and plausible. The data had never been asked. Customer knowledge was scattered across call recordings, support tickets, chat, email threads, external feedback and personal conversations, and our CRM couldn't tell us how long projects actually took or how quickly we responded.
TaskReplace the story with evidence
Before anyone designed a new screen, we needed to know where customers really got stuck, how often, and why. Not the loudest example. The pattern.
ActionOne warehouse, one classification, one brief
I pulled every customer touchpoint since the company was founded into one data warehouse: Gong, Zendesk, Intercom, email threads, external feedback, personal conversations. I built the classification system myself, categorized all of it, and ran statistics on where customers actually got stuck. An analytics tool on top gave us project durations, costs and response times that Salesforce never could.
The data said three things we had underestimated:
- Projects ran far longer than we thought.
- Response times were slow and uneven.
- Almost everything that mattered happened offline, in email, where the customer could not see it.
Those three findings became the brief for a new services-ordering system in the customer portal: one transparent flow that explains every stage and treats each service on its own terms. At the end of each stage the customer gets an automatic update, designed to replace the email chase.
I led the research and the method, then taught the team to continue from it. Ofri B., a senior product designer on my team, took it from there and led the design.
ResultThe right problems, confirmed
Stakeholder sessions and user interviews told us we had hit the problems nobody had been treating. The system is in development, so we have not measured the result yet. That number belongs in this story when it exists, not before.
What travels
When a complaint repeats, the explanation everyone already agrees on is the one most worth testing. The customer had told us everything. It just lived in six systems that never spoke to each other.