Customer journey analytics for utilities that counts every customer who had to come back — and costs it
See every customer who had to come back, and fix the step that sent them.
See one case, screen by screen ↓A customer who asks for a new connection, moves home, waits for a deposit refund or pays to be reconnected crosses the website, the app, the virtual assistant, the phone line, email, a customer centre and four or five back-office systems before the job is done. Web analytics stops at “application submitted”. The contact centre counts calls by reason. Operations counts orders by status. None of them share a request key, so no report shows one request end to end — and the customer’s two weeks of chasing appear in none of them.
The happiness survey arrives after the job is done and asks about the last visit. A customer who called four times and then met a friendly technician answers “happy”. Meanwhile every repeat contact costs money and satisfaction. In some markets experience has a price too: Ofwat’s C-MeX ties water companies’ revenue to customer experience, and in Great Britain a network operator that misses an electricity connection timescale owes the customer a guaranteed-standard payment.
Estimated split for one quarterly review of one family of journeys, in working days, by hand and with the solution.
Seven agents stitch every request across channels, count each come-back and draft the fix; the journey owner approves.
Lumina Utilities, a fictional electricity and water utility. The quarter is Q3 2026, 1 July to 30 September, stitched on Tuesday 13 October at 06:10. Leila Mansour owns new connections, move-in, move-out and refunds. Here is her morning, screen by screen, in the working solution.
The stitcher joined 4,126,300 web and app events, 61,240 contacts on five channels, 92,410 order milestones and 5,880 survey answers into 43,550 customer requests. The journey map draws seven journeys as lanes across four stages — learning about the service, starting the request, request in progress, finishing the service. A lane’s width is the customers still on track; a bubble is where they had to come back.
“→ 43,550 requests · 31,078 come-backs · 1 new hotspot”
Leila switches the map to Survey says. Every journey scores 63–86% happy; new power connection scores 81%. The lanes are smooth and green. “The survey says 81% happy. My phones say something else.”
“What the survey says. Every journey scores 63–86% happy. New power connection: 81%.”
Switched back to Measured effort, the new-connection lane runs 2,418 eligibility checks, 2,301 applications, 1,962 uploads, 1,804 fees paid and 1,690 inspections passed — then only 1,104 energised in 5 days. 586 are stuck, and the bubble reads 3.4 contacts per customer. The journey costs $18.40 per request, $24K of it in come-back contacts.
“586 customers waited for power and came back 3.4 times each — up from 1.9 in Q2.”
The hotspot finder ranks eight steps where customers come back 1.5 times or more, or where 10% of a journey is lost. HS-01, inspection passed → energised, is new this quarter: 586 customers, 3.4 contacts each, 1,992 come-backs, $14.9K. Its trend runs Q1 1.6, Q2 1.9, Q3 3.4. Together the hotspots hold 15,372 come-backs and $114K this quarter.
“Threshold: 1.5 contacts per customer, or 10% of the lane lost at one step”
The reason reader reads the calls, chats and emails behind the come-backs and groups them, keeping quotes with their source. “When will my power be on?” — 57%, 1,138 contacts. “The inspector passed it — the app shows nothing” — 17%. “My builder is waiting to hand over” — 13%. “The callback I was promised never came” — 8%. The come-back chain shows calls peaking around day 6, past the 5-working-day target.
Limit: no quote without its source; a reason is never inferred from silence.
The milestone mapper reads why each of the 586 orders waited. 375 (64%) waited for the weekly manual energisation release, Tuesdays at 07:00. 170 (29%) were also held for a meter allocation. 41 (7%) had an access problem or the customer was not home. Leila re-checks the hotspot and the agents run again: 96 new contacts since 06:10, same top reason, 12 more orders released this morning, 9 still waiting.
“The app shows “Inspection complete” with no next date, so the customer has nothing to wait for but the phone.”
Request CN-26-08817. Omar’s inspection passed on Wed 19 Aug. He chatted on Fri 21 Aug — the virtual assistant could not see connection dates and suggested calling — called on 22 and 26 Aug, was promised a callback that never came, emailed on 27 Aug and walked into the Cedar Point centre on 31 Aug. His order missed the 25 Aug release for want of a meter and went out on 1 Sep; power came on 3 Sep. Survey: Happy, 4 of 5. Measured: 5 contacts, 15 days from inspection to power.
From his call on 22 Aug: “The inspector passed it on Wednesday. Nobody can tell me when the power goes on.”
The calculator costs every contact by channel, at the rates set in Settings: phone $8.90, chat or assistant $2.60, email $4.10, centre visit $14.60 — $7.40 blended on Q3’s 60 / 21 / 12 / 7% mix. Of $340K to serve these seven journeys this quarter, $230K is come-backs: $920K a year at this rate. Deposit refunds have the costliest come-backs: $52.8K, 79% of what that journey costs to serve.
Limit: uses only the channel costs set in Settings.
B-1: release energisation automatically when the inspection passes and fees clear; show the date in the app and by SMS. Owner: Tomás Ferreira, Connections operations. Expected outcome: 90% energised within 5 working days of inspection, and under 1.0 contacts per customer at this step. It removes about 1,400 contacts a quarter, $41.4K a year, and ranks fourth of seven in a backlog worth $285K a year.
Limit: the backlog writer never sends or commits a change.
Leila approves B-1 and it goes to Tomás Ferreira, Connections operations, to accept and set the date. The status reads “Approved · with Tomás”, and the date column reads “Owner sets”. Leila can ask for B-1 to be watched after go-live: every Monday the two outcomes are measured from the stitched journeys, and both of them hear if either misses for two weeks running.
Journey owner approves · fix owner accepts and sets the date.
The same hotspot, seen by the four people who own it — what their quarter looked like, and what it looks like now.
Seven agents stitch every request across channels, count each come-back and draft the fix; the journey owner approves.
Joins web and app events, contacts, order milestones and survey answers into one journey per customer request, matching account, premise, mobile and request number within 14 days.
Maps order and account statuses and web events to each journey’s milestones and the four journey stages, and reads why a waiting order waits — release, allocation, access.
Counts every inbound contact made while a request is open, on any channel, places it on the step the request was waiting at, and builds the effort score per step and journey.
Reads the calls, chats and emails behind the come-backs at a hotspot, names the reason in the customer’s words, groups the same reasons and keeps two quotes per reason with their source.
Every night, computes contacts per customer and loss at every step, flags steps over 1.5 contacts per customer or losing more than 10% of the lane, and compares with last quarter.
Costs every contact by channel and splits each journey’s cost per request into designed service and come-backs.
For each hotspot, drafts the smallest change that removes the reason customers come back, with its owner, a measurable expected outcome and the contacts and dollars it removes, and ranks the backlog by value.
Approves each fix; the operations owner accepts it. The agents propose; a named person decides.
Ask in plain words — or say what to watch. Every answer points to the stitched requests, the order record or the call it came from.
Why does the new-connection lane narrow after inspection?
Energisation is released by hand, once a week. Of the 586 customers who waited more than 5 working days, 375 waited for the Tuesday 07:00 release and 170 were also held for a meter. The app shows “Inspection complete” with no date, so they asked: 57% of the 1,992 contacts were “when will my power be on?”. That is 3.4 contacts per customer, up from 1.9 in Q2.
Which journey costs most in come-backs?
Deposit refund: $52.8K this quarter — 7,140 come-backs at $7.40. Bill payment and reconnection is next at $51.4K. Fixes B-2 and B-4 are accepted and remove an estimated 4,820 contacts a quarter.
Show customers the survey called happy who came back 3 or more times
In Q3, 684 customers answered “Happy” after 3 or more contacts about the same request — 96 of them in new power connection. Omar Siddiqui rated the visit 4 of 5 after 5 contacts.
Watch B-1 every week after it goes live
Done. Every Monday I will measure energised within 5 working days of inspection (target 90%) and contacts per customer at that step (target under 1.0) from the stitched journeys, and tell you and Tomás Ferreira if either misses for two weeks running.
13 screens from the working solution, on its sample data. Pick one to see it large.
What counts as a journey, a come-back and a hotspot — and who decides — are settings Lumina chooses.
| Setting | Default | Choose from |
|---|---|---|
| Journeys and their milestones | 7 journeys | Edit each journey’s milestones and its on-track target — e.g. energised within 5 working days of inspection; deposit refund paid within 30 days |
| Match keys | Account · premise · mobile · request number | Change the keys — two must match to join across premises |
| Stitching window | 14 days after the last event | Change the window |
| Counts as a come-back | Any inbound contact while the request is open | Not counted: Lumina’s planned outbound calls · survey answers · marketing |
| Cost per contact | Phone $8.90 · Chat / assistant $2.60 · Email $4.10 · Centre visit $14.60 | Change each channel’s cost; the calculator uses only these |
| Hotspot threshold | 1.5 contacts per customer, or 10% of the lane lost at one step | Change either threshold |
| Approves a fix | Journey owner | Leila Mansour (connections and moves) · Hana Ibrahim (green energy) · Marco Bianchi (billing and payments) |
| Customer names | Journey owners only | Everyone else sees masked IDs |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample data · “target” = the design goal, measured in the live solution · “estimated” = our estimate: the $285K is 9,640 come-backs a quarter removed by the seven backlog fixes × $7.40 blended contact cost × 4 quarters, and the review-days split is our estimate · “typical” = published figures: E Source, survey of 15 utilities (average utility call $9.54; satisfaction about 15% lower for each extra contact); Gartner 2023 via CX Today (62% of channel shifts high effort). Regulatory context: Ofwat C-MeX; Ofgem Guaranteed Standards of Performance for connections. Lumina Utilities, its people, customers, places and records named on this page are fictional — characters and sample data in the working solution.
It measures what customers had to do to get a service — a new connection, a move, a refund, a reconnection — across every channel and system, not just what they said in a survey. Customer Journey Analytics stitches web and app events, contacts, order milestones and survey answers into one journey per request, shows where customers drop out or come back, costs it and drafts the fix.
Every night the journey stitcher joins events and contacts to the open request they belong to, when two of account, premise, mobile or request number match within 14 days. The come-back counter counts each inbound contact while the request is open and places it on the step the request was waiting at. The hotspot finder flags steps over the threshold, and the reason reader, milestone mapper and backlog writer explain why and draft the fix.
The effort score is the number of come-backs per request, counted for every request rather than the customers who answer a survey. In the sample quarter the survey reached 13.5% of customers; the come-back count covers all 43,550 requests. Lumina’s own planned outbound calls and the survey itself are not counted.
The cost-to-serve calculator multiplies each contact by the channel cost set in Settings and splits each journey’s cost per request into designed service and come-backs. In the sample quarter, $230K of $340K across seven journeys was customers coming back.
The working solution measures seven: new power connection, move-in, move-out and final bill, deposit refund, solar sign-up, new water connection, and bill payment and reconnection. Each journey’s milestones, source statuses and on-track target are settings.
Yes. The agents have read-only access and never change a customer record or an order. The backlog writer drafts each fix; the journey owner approves it, and the fix owner accepts it and sets the date.
Only journey owners. Analysts and digital product managers see masked IDs. Every quote the reason reader keeps links to the call, chat or email it came from.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your journey definitions, order statuses, channel costs and a sample of contacts and web events — and it goes live once every quality gate has passed. We will walk you through it on your own journeys first.
Every voice read, and each recurring problem named with an owner and a fix.
1-min demoClosed-Loop Feedback ManagementNo unhappy customer left without a call back, and every recovery measured.
1-min demoComplaint Root Cause AnalysisEach complaint trend proved with system data, owned, and fixed at the cause.
1-min demoWe’ll run Customer Journey Analytics on a sample of your own contacts, web events and order milestones.