Free spreadsheet · Claudia Academy
Weekly customer service scorecard
Fill in a few counts each week; the percentages work themselves out. With the definition of every KPI, so everyone counts the same way.
Version 1.0 · reviewed 2026-10-16 · goes with lesson 5
Weekly customer service scorecard (spreadsheet)
Fill in a few counts each week; the percentages work themselves out. With the definition of every KPI, so everyone counts the same way.
How to use it
- Delete the example rows first. They are fictional and only show what a completed scorecard looks like.
- Use one row per week. Write the week as the year and week number, for example 2026-W42, and let a week run from Monday through Sunday.
- Put your baseline from the customer contact audit (Lesson 1) on the first row and write “Baseline” in the Note column.
- Read the KPI definitions tab before you start, and count the same way every week: unique requests, not individual messages or calls.
- Take the counts from your lists: new requests, resolved and first response from your audit or follow-up list; callback requests and called back from the callback list; appointment requests and appointments booked from the appointments and follow-up list.
- Always count the open follow-up at the same time, for example Monday morning before opening, including items from earlier weeks.
- Only fill in the counts. The file calculates the percentages itself; do not overwrite those columns.
- In the Note column, write anything that could explain the figures: a holiday, a promotion or advertisement, illness, an outage or a new way of working.
- Compare each week with the baseline and with the three weeks before it, and discuss the figures in the weekly 15-minute improvement meeting. Only draw conclusions after four to six weeks.
What the spreadsheet calculates for you
- % resolved = resolved ÷ new requests
- % within target time = first response within target time ÷ new requests
- Callback rate = called back ÷ callback requests / missed calls
- Booking rate = appointments booked ÷ appointment requests
- Escalation rate = escalations ÷ new requests
The formulas are in the file itself; click a cell to see them. In Google Sheets, open the file with File → Import or by dragging it into Google Drive.
The columns
| Column | What you fill in |
|---|---|
| Week | Short text |
| New requests | A number |
| Resolved | A number |
| First response within target time | A number |
| Callback requests / missed calls | A number |
| Called back | A number |
| Appointment requests | A number |
| Appointments booked | A number |
| Open follow-up (end of week) | A number |
| Escalations | A number |
| Errors / complaints about contact | A number |
| Note | Short text |
Example data (fictional)
The file contains a few made-up example rows, marked in gray. They show what a completed row looks like. Delete them before you start.
| Week | New requests | Resolved | Called back | Appointments booked |
|---|---|---|---|---|
| 2026-W34 | 60 | 39 | 6 | 16 |
| 2026-W35 | 58 | 44 | 10 | 17 |
| 2026-W36 | 63 | 51 | 11 | 20 |
| 2026-W37 | 70 | 60 | 13 | 24 |
| 2026-W38 | 66 | 57 | 12 | 21 |
The KPIs, so everyone counts the same way
- % resolved
- The share of this week's unique new requests where nothing more needs to happen on your side: the question has been answered, the appointment or order has been recorded, or the customer has said they are not interested. Counted at the fixed time you fill out the scorecard.Pitfall: Counting repeat messages or several calls from the same customer as separate requests. Or counting a request as resolved because something was sent back, while the customer is still waiting for an answer.
- % first response within target time
- The share of unique new requests that got a personal first response within the target time you set yourself, for example 30 minutes during business hours and before 10:00 AM the next business day outside them. An automatic greeting or away message does not count as a first response.Pitfall: Leaving unanswered requests out of the denominator, which makes a week with many requests left sitting look better. Or changing the target time partway through without noting it in the Note column.
- Callback rate
- The share of callback requests and missed calls that were actually called back. A number that called several times counts once. A missed call from a customer who has already been helped through another channel does not need to be included.Pitfall: Counting every missed call separately, while a customer who called three times is one request. Or counting a call attempt without contact as called back, without a fixed rule on the number of attempts.
- Booking rate
- The share of appointment requests that led to a booked appointment. If you work with orders or quotes, read order or quote instead of appointment.Pitfall: Automatically crediting a rise to your new way of working. Season, promotions, holidays and capacity (an extra employee, fewer free slots) also play a part. Note them in the Note column.
- Open follow-up
- The number of requests, callback requests and appointments that have not been resolved at the fixed counting time at the end of the week, even if they come from earlier weeks. This is a count, not a percentage.Pitfall: Counting at different times, or forgetting to include items without a follow-up date.
- Escalation rate
- The share of unique new requests that had to go to the owner or a specialist under your escalation rules, for example a complaint, an emergency or a question the team could not answer.Pitfall: Drawing conclusions from the percentage when the numbers are small: going from 2 to 3 escalations is one case. Look at the cases themselves in the improvement meeting.
- Errors and complaints about contact
- The number of times per week that something went wrong in customer contact: wrong information given, an appointment double-booked or forgotten, a customer complaining that they were not called back. Complaints about the product or service itself do not belong here.Pitfall: Keeping quiet about errors because the figure feels like a check on people. Agree that reporting an error brings no blame; otherwise the figure goes down without anything improving.
- Channel mix (optional)
- What share of unique new requests comes in through each channel: phone, WhatsApp, email or website, walk-in. There is no fixed column for this; count it once a month from the Channel or Source column in your audit or appointments list, or add columns yourself.Pitfall: Counting a customer who gets in touch about the same question through two channels twice. Count them under the channel they used first.
Privacy and retention
This scorecard only contains counts per week and should not contain any personal data: no names, phone numbers, conversation content, payment details or health information, not even in the Note column. So write “two complaints about wait time,” not who complained. The lists you count from do contain customer data: give access only to those who need them for their work, do not share them through an open link, and clear out rows you no longer need. You can keep the scorecard itself for as long as you use it for comparison. Which rules on personal data apply to your business depends on your situation; when in doubt, get advice.
Made for free by Claudia Academy (Bits Please) · claudia.bitsplease.org/en/academy