Lesson 5 of 6 · 6 min

Ask Emma: Why were Tuesday bookings down?

Ask Emma questions like "Why were Tuesday bookings down?" and check the answer.

Read the lesson

Goal: Find the cause of a KPI drop with Ask Emma and the raw screens. Fix one thing. Prove the fix with 2 weeks of data.

You need: Any paid plan. Lessons 2 and 4 of this course.

Why this matters

Every client has a bad week. What the owner remembers is how fast you found the cause. A guess wastes money. A discount for a staffing problem cuts margin and fixes nothing. A clear method finds the real cause, fixes it, and shows the owner proof.

The 4-step method

  1. Ask: use Ask Emma to narrow the drop. Which day, which hours, which service, which source?
  2. Verify: check the raw records. Read the calendar, the call list and a few transcripts.
  3. Fix one thing: change only one setting at a time. Note the date.
  4. Watch for 2 weeks: compare the same KPI with the weeks before the drop.

Why one change at a time? If you change three things, you cannot tell which one worked. Why 2 weeks? One week can be luck, weather or a holiday. Two weeks shows a pattern.

Fig 1

The 4-step method on the calendar

  1. Day 0Ask: Ask Emma narrows the day, hours, service or source
  2. Day 0Verify: calendar, call list and 3 to 5 transcripts
  3. Day 0Fix one thing and note the date
  4. Days 1 to 14Watch the same KPI
  5. Day 14Compare with the weeks before the drop
One change, one date, two weeks of proof. One week can be luck, weather or a holiday.

Reading transcripts for the cause

Transcripts tell you what callers wanted and what they got. Open Analytics → Calls and pick 3 to 5 calls from the problem day and hours. Read each one and answer three questions.

  1. What did the caller ask for? A day, a time, a service or a price?
  2. What did the AI offer? Did it offer what the caller asked for?
  3. How did the call end? Booked, transferred, or the caller hung up?

A pattern across 3 calls is strong evidence. One odd call is not. If callers ask for mornings and the AI offers only afternoons, the cause is the calendar. If the AI cannot answer a price question, the cause is the knowledge, not the calendar. Fix the knowledge in Voice Agents → Knowledge → Quick answers, and test in Talk in browser, which is free. If callers hang up during the greeting, the greeting may be too long. Shorten it and test again.

Write what you found in one line in your Operator notes. Add the call times, so you can find them again. Share one short transcript line with the owner. It makes the cause real to them.

Questions that work in Ask Emma

Ask narrow questions with a time frame. Broad questions get broad answers.

  • Weak: "Why are bookings down?"
  • Strong: "Compare Tuesday bookings for the last 3 weeks with the 4 weeks before. Split by hour."
  • Strong: "List missed calls on Tuesdays between 8 AM and noon in the last 3 weeks."
  • Strong: "Which services were booked on Tuesdays last month but not this month?"

Each answer is a lead, not a verdict. Write it down, then check it in the raw screens. If Emma gives a number, find the same number in Jobs or Analytics. If you cannot find it, do not use it.

Fig 2Example numbers

Narrow questions to Ask Emma

  1. YouCompare Tuesday bookings for the last 3 weeks with the 4 weeks before. Split by hour.
  2. Ask EmmaTuesdays averaged 38 bookings before and 26 in the last 3 weeks. Most of the drop is between 8 AM and noon.
  3. YouWhich hours on Tuesday had open slots but no bookings?
  4. Ask EmmaTuesday mornings show no open slots for one cleaner. Afternoons have open slots.
  5. noteNext: check the Week view and read 3 transcripts
Ask with a time frame and a split. Each answer is a lead, not a verdict, until the raw screens agree.

By business size

For an owner-operator, a drop of 3 jobs can be one sick day. Ask the owner first. It is often faster than any tool. For a fleet, look for a pattern by tech or crew. One tech's calendar often explains a whole day. For a multi-location client, check whether the drop is in one location or all of them. A drop in one location points to that location's settings, staff or phone number. A drop in all locations points to something shared: a workflow, a booking rule or the season.

Common causes of a one-day drop

CauseWhere to look
An old Block Time break or time offJobs → Day view for that weekday
A closure entered on the wrong dateSettings → Brand & Locations → Business hours → Closures & Holidays
A booking rule changeCalendar Configurations → Booking Rules: Minimum Notice, Maximum Advance
Missed calls at certain hoursAnalytics → Calls, then call transcripts
A failed workflowThe workflow → History → View run history

Build it in ServiceAgent

  1. Open Ask Emma. Ask: "Compare Tuesday bookings for the last 3 weeks with the 4 weeks before."
  2. Ask: "Which hours on Tuesday had open slots but no bookings?"
  3. Open Jobs → Week view. Look at Tuesdays. Look for Block Time entries.
  4. Open Analytics → Calls. Look for missed or short calls on Tuesday mornings. Read 3 transcripts.
  5. Check Booking Rules and Closures & Holidays for recent changes.
  6. Fix the one cause you proved. Write the date in your Operator notes.
  7. Set a 2-week reminder: a Scheduled Trigger workflow with Create CRM Task, or a task you add by hand.
  8. After 2 weeks, compare Tuesday bookings with the 4 weeks before the drop.
  9. Check it works: report the result only when 2 weeks of data agree.
Fig 3Example numbers

What the Week view showed

Week view, Tuesday

8 AM to noonBreak, recurring weekly, for a cleaner who left a month ago
AfternoonOpen slots
Block Time
A break for a cleaner who left closed every Tuesday morning. Removing it was the one fix.

Worked example: Fresh Start Cleaning

Fresh Start has 12 cleaners on Core. Tuesday bookings fell. For 4 weeks before, Tuesdays averaged 38 bookings. The last 3 Tuesdays had 27, 25 and 26. The average is (27 + 25 + 26) ÷ 3 = 26. That is 12 fewer bookings each Tuesday. At an average of $150 a clean, that is 12 × $150 = $1,800 a week at risk.

Fig 4Example numbers

Fresh Start's Tuesday at risk

Average Tuesday before the drop38
−Average of the last 3 Tuesdays: (27 + 25 + 26) ÷ 326
=Fewer bookings each Tuesday12
×Average clean$150
At risk each week$1,800
A small calendar mistake can cost a full morning of revenue every week.

Emma pointed to Tuesday mornings. The Week view showed a recurring Break on one cleaner's calendar from 8 AM to noon. That cleaner left the company a month ago. Nobody removed the break. Transcripts confirmed it. Callers asked for Tuesday mornings and the AI offered only afternoons.

You removed the break on the day you found it. You noted the date. Two weeks later, Tuesdays had 36 and 37 bookings. The average is 36.5. That is close to the old 38, so you reported: "Fixed. Tuesday mornings are open again."

Fig 5Example numbers

Fresh Start Tuesday bookings

4 weeks before the drop38
3 weeks during the drop26
2 weeks after the fix36.5
Bookings came back after one old break block was removed. 36.5 is close to the old 38.

Common mistakes

  • Mistake: Telling the owner it is fixed the same day. Fix: Wait for 2 weeks of the same KPI.
  • Mistake: Changing several settings at once. Fix: Change one thing and note the date.
  • Mistake: Trusting the AI's cause without proof. Fix: Confirm it in the calendar and transcripts.
  • Mistake: Comparing with the wrong weeks. Fix: Compare with the weeks before the drop.

Do this now

Pick one KPI in your practice account or a client account. Ask Emma 2 questions about it. Verify one answer in the raw screens. Write a one-line cause, one fix and a review date 2 weeks out. Deliverable: a 3-line fix log. Time: 20 minutes.

Practice the conversation

  1. Owner Tuesdays have been dead. Should we run a deal?
  2. Operator Not yet. I found the cause first.
  3. Owner What was it?
  4. Operator An old break block for a cleaner who left. It closed Tuesday mornings.
  5. Owner So callers could not book mornings?
  6. Operator Right. The AI offered only afternoons. I removed the block today.
  7. Owner Is it fixed?
  8. Operator I will know in 2 weeks. I will compare Tuesdays with last month.
  9. Owner And if it is not?
  10. Operator Then I test the next cause. One change at a time.

Quick check

  1. You removed an old break block today. When can you tell the owner the fix worked?
    • Today, because the cause is clear
    • When `Ask Emma` says the problem is solved
    • At the end of the quarter
    • After 2 weeks, compared with the weeks before the drop
  2. Three Tuesdays had 30, 28 and 26 bookings. The 4 weeks before averaged 36. What is the average weekly drop?
    • 6
    • 8
    • 10
    • 28
  3. You suspect two causes: a break block and a booking rule. What do you do?
    • Fix both today to save time.
    • Fix one, note the date and watch 2 weeks.
    • Fix neither until the next monthly report.
    • Ask the owner to choose a discount.

ServiceAgent UniversityLearn to run a service business on autopilot. Free courses, Blueprints and certification for Service Operators.

Learn

Get certified

Community