Do you know that quiet moment on a Friday evening when the last customer email has been answered, the inbox finally shows nothing new, and you let yourself believe that the week went well? I know that feeling from more than fifteen years in customer service. I also know how often it lasts only until Monday morning, when the same customer writes again about the same parcel, the same invoice or the same login that still does not work. That second message tells me more about a desk than the count of closed tickets that people are usually proudest of. In our trade it is called your repeat contact rate, and it measures something very simple, which is the distance between an answer that looked finished to you and a problem that still feels unfinished to the customer.
You might know this especially well if you run a small webshop and answer the customer emails yourself after closing time, because then every customer who comes back lands on your own evening. So I would like to show you how to count this number on a small desk, and which habits bring it down. This autumn the bigger names in customer service happened to give us a good reason to look past the empty inbox.
Why your closed count and resolution rate look better than they are
At the end of September, Twilio published research showing that 86% of companies believe their customers are satisfied with AI-powered service, while only 51% of consumers agree. Twilio sells these tools, so I read it as a vendor survey. Still, I recognise its explanation from my own desk, because a conversation the AI kept away from a person counts as contained whether the customer got what they needed or simply gave up and stopped replying. In the data, those two very different endings look exactly the same.
A few days earlier, in Customer Service Manager Magazine, Marija Jovanović of Helpjuice, who openly says her employer sells an AI agent, used a worked example to show how one week of conversations can be reported as a 45% resolution rate or a 93% one, and the only difference between the two is how you count the customers who went quiet. In mid-September, CX Dive then reported on a CallMiner survey of 700 senior contact centre and CX decision-makers in which 76% said their automation is driven by internal goals rather than by what customers need.
Your webshop does not run a contact centre, and none of this is really about robots. A small desk also measures what is easy to see, every time a case is marked solved simply because nobody wrote back by Thursday.
Count your repeat contact rate over seven days
Your repeat contact rate is the share of customers whose case you closed and who came back about the same problem within a set window. To take an imagined week as an example, if you closed cases for forty customers and six of them wrote again about the same thing, your rate would be six out of forty, or 15%. It sits close to first contact resolution, which asks whether the problem was solved in one conversation, and to reopened tickets, which only catch the people who reply on the old thread. The repeat count also finds the customer who started a new email or simply picked up the phone.
Twilio and Jovanović both suggest a seven-day window, and Twilio calls repeat contact within seven days the cleanest proxy for something the first interaction failed to fix. Jovanović adds a detail that matters a great deal on a small desk, which is to count by customer rather than by ticket and on any channel, so that a repeat email does not quietly pass for new demand.
In practice this is a small habit for Friday afternoon. You take the week's closed cases, look at who wrote again with the same order number or the same question, and write that number next to your closed count, and the customers behind it become your real homework for the week ahead.
Confirm the fix before you close the ticket
To bring that number down, start by not guessing. When I believe a case is done, I tell the customer in plain words what I did and ask one honest question. It can be as simple as "I have resent the invoice to the address on your account, could you check that it arrived and tell me if anything still looks wrong?" It takes a moment, and it turns a closed ticket into a confirmed one. A closed ticket is a fact about your inbox, and only the customer's answer makes it a fact about the customer.
If the customer goes quiet after that, I do not take the silence as a thank you. I leave the case open for a day or two and check what I can myself, whether that is the carrier's tracking page or the payment in the bank, so that when I finally close it, I know rather than hope.
Read every comeback as a question about the first answer
Say a customer writes on Wednesday that the download link you sent on Monday still does not work. Before I answer, I read my first reply again, because the first answer often fixes the symptom and misses the cause, which here may simply be a link that expires after an hour. Asking myself what I assumed and what I promised usually shows me the gap quickly, and it is the same patient habit I use when I troubleshoot a problem I have not seen before.
A returning customer has already spent her patience on you once, so open the history before you reply and spare her from explaining everything again. Tell her plainly that you can see this is the second time and that you will stay with it until it is done, because in my experience that one sentence calms people better than any apology macro.
Write the case note for whoever opens it next
Repeat contacts hurt most when the second conversation starts from zero. That is why I write every case note for the person who might open it next week, who may be a colleague or may be me on a tired Tuesday. I keep it to the order number, what went wrong, what I did, what I promised and by when, and if you would like the longer version, I wrote more about how I set myself up in a CRM.
If you want a quiet place to practise this with your team, Biurko, the small CRM and customer service practice desk on this site, is there for exactly that, although a spreadsheet and good habits will also take you a long way.
The short version
Your closed count tells you how busy you were, while your repeat contact rate tells you how well it actually went. Keep the two numbers side by side every Friday and let the second one decide what you work on next. Ask before you close, treat silence as a question rather than a thank you, and write case notes that spare the customer from starting over, because that is how the empty inbox on Friday starts to mean what you hoped it meant.
Questions about repeat contact rate
What is a repeat contact rate?
A repeat contact rate is the share of customers who contact you again about the same problem within a set window after you closed their case. You work it out by dividing the number of returning customers by the number of customers whose cases you closed, so in an imagined week with six returns out of forty closed cases, the rate would be 15%. Twilio and Customer Service Manager Magazine both suggest a seven-day window.
How is repeat contact rate different from first contact resolution?
First contact resolution asks whether a problem was solved in a single conversation. Repeat contact rate looks back a week later and asks whether the cases you closed actually stayed solved. A team can close quickly and still see people return, so the two numbers are most useful side by side. Reopened tickets are a narrower signal still, because they only count customers who reply on the old thread.
Is every repeat contact a failure?
No. A customer who comes back with a new question, or writes to say the parcel finally arrived, is not a repeat contact in this sense. Count only the people who return because the same problem is still there, and match them by customer rather than by ticket, so that someone who switches from email to phone is still counted once.
Do I need special software to track repeat contacts?
No. On a small desk, a weekly look at your closed cases and a simple tally of who came back about the same problem within seven days is enough to start. A CRM that keeps each customer's history in one place makes the Friday check quicker, but the habit of confirming the fix before you close matters more than the tool.
Sources
- Twilio (vendor research), "86% of brands think their AI works. Only 51% of customers agree.", Jesse Sumrak, September 29, 2026. https://www.twilio.com/en-us/blog/insights/improve-ai-satisfaction
- Customer Service Manager Magazine, "How to Read an AI Resolution Rate Before You Believe It", Marija Jovanović (Helpjuice, a vendor), September 24, 2026. https://customerservicemanager.com/how-to-read-an-ai-resolution-rate-before-you-believe-it/
- CX Dive, "Most companies underuse CX data to guide automation efforts", Matt Vartabedian, September 15, 2026 (reporting a CallMiner vendor survey). https://www.customerexperiencedive.com/news/most-companies-underuse-cx-data-guide-automation-efforts/830336/
How can I help you today?
If you would like help setting this up on your own desk, in Polish or English, write to me at info@kozdrach.com and I will reply within 24 hours. If you would rather talk it through, you can also book a quiet 1:1 session with me of 30 minutes, 45 minutes or 60 minutes, the same sessions you will find on my homepage.