Key takeaways
- Phone systems record that a call happened, not how it ended; outcome tags fill that gap.
- Tag two things on every call: who called (new, existing, supplier or partner, junk) and how it ended, from a list of about ten outcomes.
- Tag at the end of each call, one outcome per call, and read every "Other" at a weekly review.
- Five reports come out of the tags: new-caller booking rate, overdue promises, reasons for not booking, junk share and outcomes by hour.
- Missed calls need tags too, one for each return attempt, because most people let unknown numbers go to voicemail[2].
Your call log knows the time, not the result
Ask an office manager how many new-customer calls turned into bookings last month and the honest answer is usually an estimate. The phone system can say how many calls came in, how long they lasted and how many were missed. It can't say whether the caller booked, wanted a quote, was a supplier chasing an invoice or was a robocall. Those are the questions the business actually runs on, and they are the ones nobody can answer on a Monday morning.
This matters because finding and keeping customers is the hard part of running a small business. In the Federal Reserve Banks' 2024 Small Business Credit Survey, reaching customers and growing sales was the operational challenge employer firms reported most often, at 57%, ahead of hiring or retaining qualified staff at 51%[3]. The phone is where a lot of that reaching happens, and without outcome tags you can't see where it leaks.
Writing down what happened on a call isn't new admin either. In O*NET's ratings of customer service work, keeping records of customer interactions, including the actions taken, scores 88 out of 100 for importance, just behind talking to customers in the first place[1]. Disposition codes, sometimes called call outcome tags, make that record short and consistent enough to count.
A short code list for a small office
Tag two things on every call, separately. Caller type says who it was: new customer, existing customer, supplier or partner, or junk. Outcome says how the call ended. Keeping them apart is what makes the reports work, because "booked" means something different for a first-time caller than for a regular moving their Thursday appointment.
| Code | Use when | Next step and owner |
|---|---|---|
| Booked | An appointment, visit or job went into the calendar | None, or a confirmation from the front desk |
| Quote or estimate | The caller needs a price that someone has to prepare | Estimator, with a due date |
| Callback due | Someone promised to call back | Named person, with a time window |
| Answered, no booking needed | A question was answered and nothing more was needed | None |
| Not booked: reason | A potential customer chose not to book; record why | Manager reviews the reasons weekly |
| Not a fit | Outside your area, a service you don't offer, or a job you don't take | None; count them for marketing |
| Transferred | Passed live to another person | The person who took the call |
| Message taken | The person wanted was unavailable and a message was left | The recipient, by end of day |
| Junk | Robocall, sales pitch or wrong number | None |
| No contact | A return call that reached nobody; note whether a voicemail was left | Same person, second attempt |
Ten codes is about the limit for people tagging calls between other jobs. Past that, staff start picking whichever code is nearest the top of the list, and the reports drift without anyone noticing. If you need more detail, put it in a reason field under one code, the way "Not booked" carries price, timing, went elsewhere or never responded. Those reasons deserve their own short list; lost lead reasons has one to start from.
Rename the codes to match how your team talks. A salon might split Booked into New client and Rebooked; a plumbing office might split it into Emergency dispatched and Scheduled visit. What matters is that every code answers one question, "what happens next?", and that two people tagging the same call would pick the same code. Test that before the trial: give two staff members the same ten call summaries and compare their tags. Wherever they disagree, the definition needs one more sentence.
Rules that keep the tags honest
- Tag at the end of the call, before the next one. Tags added at the end of the day are guesses dressed up as data.
- One outcome per call. If a call did two things, tag the one that matters most to revenue and mention the other in the notes.
- "Other" is allowed, but it needs a note, and every "Other" is read at the weekly review. Three of the same note means a new code, or two codes that should merge.
- Open outcomes need an owner and a date. A callback or quote with neither is a promise nobody owns.
- Tag the attempts, not just the calls. Each return call on a missed call gets its own outcome.
The overhead is small if the rules are followed. Example: an office taking 50 calls a day that spends ten seconds tagging each one uses roughly three hours of staff time a month. At the BLS average of $36.73 an hour in employer cost for office and administrative support workers[4], that is about $110 a month, for a record that tells you where new business comes from and where it slips away.
- System
- Your team
01System
Call ends
Notes, or the call summary, are on screen.
02Your team
Tag the caller type
New, existing, supplier or partner, or junk.
03Your team
Tag the outcome
One code from the list, plus a reason if it wasn't booked.
04Your team
Set owner and due date
Only for open outcomes: callback due, quote, message taken.
05System
Weekly report
Counts by code, overdue items and every Other note.
Example flow. In a spreadsheet the tags are two drop-down columns; in a CRM they are two required fields on the call record.
Five reports the tags make possible
Once the tags have been consistent for two or three weeks, five simple counts answer most of the questions owners ask about their phones. None of them needs software beyond a spreadsheet filter, and ten minutes on Monday morning is enough to read them as a team and agree one change for the week.
- New-caller booking rate: new-customer calls tagged Booked, divided by all new-customer calls. This is the number to watch when you change scripts, staff or opening hours.
- Overdue promises: every Callback due, Quote or Message taken past its date. Read this one daily, not weekly.
- Why people didn't book: the Not booked reasons, counted. Price, timing and "went elsewhere" each point to a different fix.
- Junk share: calls tagged Junk as a share of all calls, by hour. If it's high, tighten screening; if it's near zero, check whether real junk is hiding under Other.
- Outcomes by hour of day: which hours produce bookings and which produce messages. That is your staffing schedule, drawn by your own callers.
calls answered or returned
from new customers
tagged Booked
quotes still open at month end
Example figures for illustration, not benchmarks.
Example: the month above gives a new-caller booking rate of 31 out of 74, about 42%. If the Not booked reasons show 15 callers said "timing", the fix is probably earlier appointments or a callback when the next opening appears, not a bigger advertising budget. That is the point of tagging: a number that tells you which lever to pull. The same tags are what a fair trial of any new phone setup needs; the two-week test for an AI receptionist starts with exactly this kind of log.
Tags for calls nobody answered
Missed calls are where tagging pays off fastest, because the outcome isn't settled when the phone stops ringing. Tag every return call as reached, voicemail left or no answer. The voicemail matters. In Pew's survey, 67% of US adults said their general practice is to let an unknown number go to voicemail and check the message, and only 19% said they generally answer[2]. A return call that leaves a clear message with a reason to call back is a real attempt; one that hangs up at the beep isn't.
If some of your calls come through Google's Local Services Ads, the tags also show how responsive you look from the outside: Google says missed calls may count against the responsiveness it uses in ranking[5]. And tag junk honestly instead of deleting it from the log. Robocalls still accounted for 1,601,611 of the 2,618,077 Do Not Call complaints the FTC received in fiscal 2025[6], so a junk count that suddenly jumps is worth noticing, and it is the evidence you need when deciding whether stricter screening is worth the risk to real callers.
Where the tags live
A spreadsheet works for a single location: one row per call, drop-down columns for caller type and outcome, a reason column, an owner and a due date. A CRM works better once more than two or three people answer the phone, because the tags sit on the contact record and open items show up as tasks. The call log template has the columns ready to copy.
If an SI Employee answers some of your calls, each one arrives as an email summary, with the recording and transcript in the account, so whoever tags it is reading a few lines rather than listening back. Call data can also be sent by webhook, through Zapier, Make or n8n, into the sheet or CRM where your tag columns live, so software-answered and staff-answered calls end up in one report. A person still applies the outcome tag, because only the team knows whether the quote went out or the customer turned up. The real estate team example follows lead calls through to the agent who owns them, and home service companies can see how outcome counts feed a CSR scorecard.
Start next Monday
Print the list
Put the ten codes and the caller types on one page by every phone. Rename codes to your own words now, not halfway through the trial.
Run two weeks
Everyone tags every call, including return calls on missed ones. Nobody changes the list mid-trial, however tempting.
Read the Others
At the end of week two, read every Other note. Add a code only if the same note appears three times.
Prune
Remove any code used fewer than twice in two weeks and fold it into its nearest neighbor.
Lock it and report
Freeze the list for a quarter and send the five reports every Monday. New hires learn the codes in the first week of their phone training.
Give the trial an honest start by telling the team why it's happening: not to check up on anyone, but to find out which calls turn into work. Reports that are used to blame people get gamed within a week. Reports that are used to fix scripts, hours and follow-up get more accurate over time, because the people tagging can see the point. If you are weighing what kind of help to add at the front desk once the numbers are in, what an SI Employee is explains the option in plain terms.
Frequently asked questions
How many call disposition codes should a small business use?
About eight to ten outcomes, plus a separate caller type. Fewer and the reports are too coarse to act on; more and people pick the nearest code instead of the right one.
What's the difference between a call log and disposition codes?
A call log records that a call happened: time, number, duration. A disposition code records how it ended. You need both, but only the codes tell you whether calls are turning into business.
Should missed calls get a disposition code?
Yes, and so should each attempt to return them: reached, voicemail left or no answer. That is how you see whether missed calls are being recovered or quietly dropped.
Who should tag the calls?
Whoever handled the call, straight after it. For calls taken by an answering service or software, the person who reads the summary and does the follow-up applies the tag.
Can software tag call outcomes automatically?
Some systems suggest an outcome from the conversation, but treat that as a draft. A person who knows what happened next should confirm the tag, especially for quotes and callbacks that are still open.
References
O*NET OnLine (U.S. Department of Labor, Employment and Training Administration) (2025). 43-4051.00 - Customer Service Representatives (opens in a new tab) ↑
Pew Research Center (2020). Most Americans don't answer cellphone calls from unknown numbers (opens in a new tab) ↑
Federal Reserve Banks (Small Business Credit Survey) (2025). 2025 Report on Employer Firms: Findings from the 2024 Small Business Credit Survey (opens in a new tab) ↑
U.S. Bureau of Labor Statistics (2026). Employer Costs for Employee Compensation, Table 4: private industry workers by occupational and industry group, June 2026 (opens in a new tab) ↑
Google (2026). About ad rankings (Local Services Ads Help) (opens in a new tab) ↑
Federal Trade Commission (2025). National Do Not Call Registry Data Book, Fiscal Year 2025 (opens in a new tab) ↑