Auto Attendant Reporting: Call Metrics Worth Tracking

Why reporting matters even with three phones

A small business with one number and two or three people answering it does not need a call center dashboard. It needs answers to four plain questions: how many people called, how many reached a person, how many gave up, and how many called when nobody was there. Everything else is detail.

The reason to track these at all is that missed calls are invisible. A caller who hangs up during the greeting leaves no message and no complaint. If you only look at voicemails, you see the people patient enough to leave one, not the ones who went to a competitor. A five-minute weekly look at a handful of numbers is enough to spot a menu that confuses people, a lunch hour nobody covers, or a closed-hours message that is losing work. The page on missed calls covers what happens to the caller on the other end.

This article does three things: defines the five numbers worth tracking, explains where the published benchmarks come from (and why most of them are not your targets), and gives you a one-page weekly worksheet.

The five numbers worth tracking

Two supporting numbers make the five more useful:

• Messages left. Of the calls that went to voicemail, how many left a message. A low share suggests people do not trust that anyone will call back.
• Callback time. For messages left during business hours, the time from the voicemail arriving to the first return call.

Keep the categories mutually exclusive for in-hours calls: answered + voicemail + abandoned should add up to the in-hours total. Count after-hours calls separately, because nobody was expected to answer them.

Where the famous benchmarks come from

If you search for "call abandonment rate", you will find a few numbers repeated everywhere. They are real, but they come from a specific kind of operation.

• New York State's call center standards define abandonment rate as "the percentage of customers who hang up before their call is answered", say it "should ideally be kept under 5%", and give "a common SLA target" of answering 80% of calls within 20-30 seconds ^[1]. The page adds that the target can vary with the agency and the complexity of the question.
• The U.S. General Services Administration's contact center guidance, published in 2019, suggests that no more than 3% of calls be disconnected by the customer before a representative answers, and cites a high-performer benchmark of at least 80% of phone calls answered within 20 seconds ^[2].
• The City of New Orleans publishes its 311 call abandonment rate as a performance measure, with a 2018 goal of "not more than 7 percent" (the 2016 goal was 8 percent) ^[3]. The quarterly figures, from early 2014 to mid-2018, range from about 1.5% to about 7%.

All three describe government contact centers: a queue that holds callers in line, many agents logged in at once, and staff whose whole job is answering the phone. In that setting, a caller who hangs up has usually been waiting in a queue, and the 80/20 figure measures how fast the next free agent picks up.

A two-to-five person business has none of that. There is no queue; the phone rings one person, then another, while the same people serve customers, drive between jobs or work at a bench. Do not adopt 3%, 5%, 8% or 80/20 as targets. Use them as context for what the terms mean, and set your targets from your own first month of data.

Option choice rate: which keys callers press

If your number has a phone menu, the split of key presses tells you whether the menu matches why people call.

Option choice rate = calls that chose an option ÷ calls that reached the menu.

Larger systems report this directly. Microsoft's Teams auto attendant report, for example, counts "caller actions (key presses, voice commands)", the average number of caller actions per call, and a final result for each call, such as "Terminated (No Caller Action)", "Terminated (With Caller Action)", "Transferred (Voicemail)" and "Transferred (User)" ^[4]. Those two "Terminated" lines are the useful distinction: one is a caller who hung up without pressing anything, the other a caller who pressed something and still hung up.

What to look for, whatever tool you use:

1. One option takes most of the calls. Move it to key 1, or consider sending all calls there with no menu at all.
2. An option is almost never chosen. Remove it. Every option you cut shortens the greeting for everyone.
3. Many callers press nothing. The greeting may be too long, or the options do not describe what callers want. Rewrite the labels in the caller's words, not your department names.
4. Callers press the "anything else" option most. Your option labels are missing a common reason for calling.

Templates for short menus are in call flow templates, and the evidence on menu length and wording is collected in phone menu design evidence.

Abandonment: definition, formula and what to count

The general definition is simple: a caller who hangs up before being answered ^[1]. In a queue-based system, Microsoft defines an abandoned queue call as one where the caller disconnected "before a representative (agent) answered or before Call Timeout occurred", and reports calls received and abandoned per hour ^[4].

For a small business without a queue, a practical version is:

Abandonment rate = in-hours calls where the caller hung up before a person answered and before voicemail picked up ÷ in-hours calls received.

Decide in advance how you will treat three edge cases, write the rule down, and keep it the same every week:

• Very short calls. You may choose to exclude calls that end within a few seconds as likely misdials. That is a local convention, not a standard; if you use one, pick a cut-off and keep it.
• Hang-ups at voicemail. A caller who reaches voicemail and hangs up without a message is not abandoned by the formula above; count them under "went to voicemail" and track "messages left" separately.
• Repeat callers. If the same number calls three times in ten minutes and gets through on the third, that is one customer and two abandoned calls. Note it; it usually means the ring time is too short or the wrong person is first in line.

Where abandonment happens matters as much as how often. A caller lost during the greeting points to the menu. A caller lost while the phones ring points to who is first in line and for how long.

Answer within X seconds, scaled down

The 80/20 idea (80% of calls answered within 20 seconds) measures speed to a live person ^[2]. The useful part for a small team is the shape of the measure, not the numbers.

A scaled-down version: share of in-hours calls answered by the first or second person in line. If your setup rings person A, then person B, then person C, track how often the call gets past B. If calls regularly reach the third phone or voicemail during business hours, the first two people are not available often enough at those times.

Ring length is a trade-off: too short and the person barely has time to pick up, too long and the caller waits. Change one thing at a time and compare two weeks before and after.

Voicemail rate and callback time

No public standard defines a target voicemail rate for a small business; treat both numbers below as your own baseline.

Voicemail rate = in-hours calls that reached voicemail ÷ in-hours calls received.

Callback time = time from the voicemail arriving to the first return call, for in-hours messages. Track the median, or simply the share returned the same business day.

A rising in-hours voicemail rate usually has an ordinary cause: someone who covers the phone is away, or a busy hour has nobody assigned. A slow callback undoes the value of the message; callers who leave one have done what you asked.

After-hours calls are different. Every one of them goes to the closed-hours message, so the useful questions are: how many arrive in the hour before you open or after you close, and how many leave a message? Clusters at the edges of your hours suggest moving the hours, or adding a line to the closed message about when you reopen.

A one-page weekly tracking worksheet

Copy this into a spreadsheet or onto paper. Fill it in at the same time each week; it takes about five minutes once you know where each number comes from.

Then compute three ratios:

• Answered rate = answered ÷ in-hours calls
• Voicemail rate = went to voicemail ÷ in-hours calls
• Abandonment rate = abandoned ÷ in-hours calls

The five-minute review:

1. Compare this week's three ratios with the average of your first four weeks. Ignore small moves; a business with 40 calls a week will see a few points of noise.
2. Read the "notes" column from last week. Did the change you made move the number you meant it to?
3. Look at the times of abandoned and voicemail calls. Are they bunched at one hour or day?
4. Check that every in-hours message got a callback.
5. Write down at most one change for next week: a ring order, a ring length, an option label, the hours.

Where the numbers come from. Use whatever your phone service shows and fill the gaps by hand. With Auto Attendant, the dashboard shows calls and minutes for the current month and a list of recent calls, and voicemails arrive by email as MP3 files followed by a transcript, so the voicemail and messages-left columns can be tallied from your inbox. The dashboard does not break calls down by menu option or report abandonment, so those columns are a manual count or a sample. A week of tally marks by the person who answers most calls is enough to see a pattern.

FAQ

What is a good abandonment rate for a small business?

There is no published benchmark for businesses of two to five people. The figures usually quoted (under 5% for New York State's call center standards, no more than 3% in GSA's 2019 guidance, not more than 7% for the New Orleans 311 line in 2018) come from government contact centers with queues and many agents ^[1][2][3]. Measure your own rate for a month, then aim to bring it down and keep it stable.

Do I need call recording to report on calls?

No. Every metric in this article is a count or a time. A recording tells you what was said, not how many callers got through. Auto Attendant does not record calls.

How often should I review the numbers?

Weekly for the five-minute worksheet, monthly for trends. Smaller businesses see more week-to-week noise, so judge a change over at least two weeks before keeping or reversing it.

What if one option gets almost all calls?

Make it option 1, or remove the menu so all calls go straight to the people who handle them. The trade-offs between a menu, a ring group and a queue are covered in call queue, ring group or auto attendant.

Auto Attendant gives a US or Canadian business number a short phone menu that rings up to three phones in turn and sends voicemails by email, with monthly call and minute counts in the dashboard.