Measuring the Funnel

First-timer cohorts, the second-visit rate, 30/60/90-day return, regulars, visit frequency, and ambassadors: the central chapter.

If you read only one chapter of this book twice, make it this one. Everything before it has been preparation, and everything after it is a variation. The question at the center of a honky tonk’s PR is whether the people who come once come back. Not how many people came, and not how many saw a post. Whether a stranger became a regular. This chapter shows how to count that, with a pencil if you have to. That one question, whether strangers become regulars, is the closest thing a honky tonk has to a single measure of health.

The method is called cohort tracking, and the idea is simple. A cohort is a group of people who started at the same time: everyone who came for the first time in March. You follow that group forward. How many came back within thirty days? Within sixty? How many came three times? How many brought a friend? Then you do the same for April and compare. Each cohort is a small experiment, and the rates are the results. The funnel from Chapter 3 becomes a set of numbers you can watch and improve. Cohort tracking sounds technical, and it isn’t. It’s a way of keeping track of groups of people over time, and it works on index cards as well as in a spreadsheet.

A cohort beats a running total because it keeps the groups apart. If you lump everyone together, a flood of new people one month hides a leak among the old ones, and a quiet month hides a success. Cohorts show what happened to each group, and they let you tell whether a change you made, a new welcome or a new lesson, helped the people who arrived after it. This is the measurement that tells you whether the tactics in Part Two are working. By the end of the chapter you’ll have a cohort sheet and a way to fill it in. It’s the central chapter of the book, so it goes slowly and shows every step.

PR TACTIC: The Cohort Card

What it is: Give every first-timer a small card with the date, and stamp or initial it each time they return, so you can count visits without recording names.

Why it works: The card lets you follow a group over weeks with no software, and it works at the door, the bar, or the front of a class.

Try this: Print or hand-write fifty cards this week, with a date line and six boxes, and hand one to every first-timer on your key night.

25.1 Cohorts: Following a Group, Not a Pile

A cohort is everyone who first came during one period, usually a week or a month. Choose the period to fit your numbers. If you get ten first-timers a month, use a month. If you get a hundred a week, use a week. Whatever you pick, stick with it. For each cohort, the first number is its size: how many first-timers arrived. After that you count, at fixed points, how many of them came back. Because you follow the same people forward, you’re always comparing a group with its own past, not with a mix of newcomers and old hands. The difference between a cohort and a pile matters. A pile is everyone who came this month, new and old mixed together. If the pile grows, you can’t tell whether new people are sticking or old regulars are coming more often. A cohort separates them, so you can see exactly what happened to the people who were new.

There’s a subtlety worth stating. Cohorts take time to mature. The March cohort’s thirty-day number is known at the end of April, its sixty-day number at the end of May, and its ninety-day number at the end of June. Don’t compare a young cohort’s ninety-day rate with an old one’s, because the young one hasn’t had time. Compare like with like: thirty-day with thirty-day, across cohorts. A sheet with the cohorts as rows and the time points as columns makes this clear. Empty cells are honest. They just mean the clock hasn’t run out yet. A practical layout is one row per cohort, with columns for cohort size, back within 30 days, 60 days, and 90 days, regulars, average extra visits, and Ambassadors. Fill in each cell when its date arrives. After six months you will have six rows, and the pattern will be easy to see.

The rule for who counts as a first-timer must be fixed, too. A first-timer is someone who hasn’t been to this night, or this room, before. You’ll sometimes be wrong, because people forget, and that’s acceptable if you’re wrong consistently. A person who comes back after a year away is a “returning lapsed” guest, and you might track that separately. Write the definition at the top of the sheet. Don’t shift it midstream. A cohort sheet is a long-term record, and its value comes from the definitions staying the same. If you change the definition for a good reason, draw a line across the sheet, write the date and the new rule, and start fresh below. Comparing rows above and below the line is fine, as long as everyone knows the line is there. A note at the bottom of each row about anything unusual that month, such as a new teacher, a holiday, or a flood, will save confusion later.

25.2 The Numbers: Second Visit, 30/60/90, Regulars, Frequency, Ambassadors

Seven numbers cover the funnel. The first is cohort size: first-timers in the period. The second is the second-visit rate, also called the thirty-day return: the share of the cohort who came at least once more within thirty days of their first visit. The third and fourth are the sixty-day and ninety-day returns: the share who came back at least once within sixty and ninety days. These are cumulative, so each is at least as large as the one before. Together they show how fast people drift away. A big drop in the first thirty days and a flat line afterward means your leak is early, which points at Chapter 10. The third and fourth are the sixty-day and ninety-day returns, which use the same method with longer windows. They show whether people who came back once kept coming. If the thirty-day rate is healthy but the ninety-day rate is low, people are trying a second night and then drifting away, which points to Chapters 9 and 10: rituals, names, and reasons to come back.

The fifth is the regulars count. This book’s working definition is a person who has visited at least three times, including the first, within sixty days of the first visit. The share of the cohort who qualify is the regular rate. That’s a rule of thumb. If your night is monthly rather than weekly, three visits in sixty days is impossible, so change it: two visits in ninety days, or whatever fits your rhythm, and write it down. The sixth is visit frequency: among those who return, the average number of additional visits in ninety days. It says how deep the habit goes, not just whether it exists. You can change the regular rule to fit your night, as Chapter 3 explained, but use the same rule for every cohort. Frequency matters because two rooms can have the same number of regulars with very different habits. Regulars who come every week are a different foundation from regulars who come every third week.

The seventh, ambassadors, closes the loop. Count the regulars who brought someone who then showed up, which you learn from the how’d-you-hear question, from the guest’s own mention, or from a card handed to a friend. Ambassadors per cohort, or guests brought per regular, tells you whether the community is recruiting. Between these numbers and the funnel stages of Chapter 3, you have a full picture: how many came, how many stayed, how many became part of the place, and how many brought others. Seven numbers, one sheet, one page per night. Each of the seven has a natural question attached. Size: are enough people finding us? Thirty-day return: was the first night good? Sixty and ninety: is there a reason to keep coming? Regulars: has it become a habit? Frequency: how strong is the habit? Ambassadors: is the community recruiting?

THE SEVEN FUNNEL NUMBERS ON ONE CARD

1. Cohort size: first-timers in the period.

2. Thirty-day return: came back within 30 days ÷ cohort size.

3. Sixty-day return: came back within 60 days ÷ cohort size.

4. Ninety-day return: came back within 90 days ÷ cohort size.

5. Regulars: met the regular rule ÷ cohort size.

6. Frequency: average extra visits in 90 days among those who returned.

7. Ambassadors: regulars who brought someone who showed up.

Write the counts beside every rate, such as 21 of 42 (50%).

25.3 How to Count Without Software

The hard part isn’t the arithmetic. It’s knowing who’s who. The cheapest method is the cohort card at the top of this chapter: a card with the first-visit date and a few boxes. The guest keeps it, and the host or door person stamps or initials a box on each return. When a guest hands you a card with a stamp, you know their visit count without ever writing a name. At the end of the period, you count how many stamped cards you’ve seen. A card that’s lost is a lost data point, so expect to undercount, and count the same way every time. Stamps work too. A small stamp on the hand or a punch on a card for each visit lets the host see at a glance whether someone is new or returning, without keeping any list. At the end of each night, tally the stamps by visit number.

A second method is a sign-in sheet or a list of first names with permission, kept by the host. It’s more accurate for classes, where people are used to signing in. It’s a little more intrusive, so say why you’re asking and keep the sheet private. A third is the point-of-sale or ticketing system, if it can tie visits to a person who agreed to it, such as a loyalty feature. This is the most accurate and the most subject to privacy rules. A fourth is the host’s memory and a notebook, which works surprisingly well in a small room where the host knows most faces, but doesn’t scale. Pick the lightest method that works for your size. Classes have an advantage here, because a teacher sees the same faces each week and a roll call is normal. A teacher who writes first names and dates in a notebook already has a cohort sheet. All that’s left is to count each series as a cohort and fill in the columns.

Whichever you choose, be honest about the error. Cards get lost. Hosts forget a face. Someone comes with a friend who never took a card. These errors tend to undercount returns, since a returner who didn’t show a card looks like a first-timer next time. If your method undercounts by ten percent, your return rates will be ten percent low, consistently, and the comparison across cohorts still works. That’s why consistency matters more than precision here. And be careful with data you keep: first names and a date are plenty. Chapter 35 explains why you shouldn’t collect more than you need. Count the error, too. If you know that about one card in five gets lost, say so when you report, and keep the method the same so the error stays about the same size. A consistent undercount still shows the direction of change. If you change methods, run the old and the new side by side for a month so you can see how much the change itself moves the numbers.

25.4 Reading the Funnel: Where Is the Leak?

With a few cohorts on the sheet, patterns appear. Start with the biggest drop in real people, not in percentages, the lesson from Chapter 3. If most first-timers never return, the leak is at the second visit and the fix lives in Chapters 10 and 11: the welcome, the first win, the return path. If most come back once but few become regulars, the leak is at the habit stage, and the fix is a ritual, a role, a reason to come weekly, as in Chapters 7 and 9. If regulars are plentiful but ambassadors are few, the community isn’t recruiting, and the fix is an invitation. Each pattern points at a different part of the book. Look at the sixty- and ninety-day columns too. A room whose thirty-day rate rises while its ninety-day rate stays flat has improved the first night without improving the reason to stay. A room whose ninety-day rate rises with a steady thirty-day rate has built a stronger habit among the people who already came back.

Then watch the change over cohorts. A single cohort tells you where you are, and a series tells you whether what you did helped. If the second-visit rate rises across three cohorts after you introduce a welcome, that’s a promising sign. If it rises one month and falls the next, it may be noise: with a cohort of twenty, three people change the rate by fifteen points. Look at the number of people as well as the percentage. Use the baseline thinking from Chapter 23: what’s the normal swing across your cohorts? A change inside the swing isn’t a result yet. Small cohorts need care. A cohort of twelve that has six return looks like 50%, and one of twelve with four returning looks like 33%, but the difference is two people. When cohorts are small, combine two or three months into one row before drawing conclusions, and always write the counts next to the rates.

Be especially careful with the temptation to count only the survivors. A room that tracks regulars but not first-timers sees only its success. It doesn’t see the hundred people who came once and left. That’s why the cohort sheet starts from the first-timers, not from the regulars. Whatever the regulars say, whatever the room feels like, the leak is in the people who never became them. If the number of first-timers is large and the number of regulars is small, don’t buy more reach until you’ve fixed the leak, which is exactly the argument of Chapter 1. Keep track of lost people too, even roughly. A short note of why people said they didn’t come back, gathered by the listening methods in Chapter 33, is the most direct clue to fixing the leak.

THREE COHORTS, A WORKED EXAMPLE

• January. First-timers: 40. Back in 30 days: 12 (30%). Back in 60 days: 16 (40%). Back in 90 days: 18 (45%). Regulars (3+ visits in 60 days): 7 (17.5%). Ambassadors: 3.
• February. First-timers: 44. Back in 30 days: 19 (43%). Back in 60 days: 24 (55%). Back in 90 days: 26 (59%). Regulars (3+ visits in 60 days): 12 (27%). Ambassadors: 5.
• March. First-timers: 38. Back in 30 days: 15 (39%). Back in 60 days: 20 (53%). Back in 90 days: not yet. Regulars (3+ visits in 60 days): 10 (26%). Ambassadors: not yet.

Illustrative figures. Percentages are rounded and each is the count divided by that cohort’s first-timers. A blank means the clock has not run out yet.

WHEN COHORTS ARE SMALL

If a cohort has fewer than about twenty people, a change of one or two people moves the rate by five to ten points. Three habits help:

Combine: add two or three months into one cohort row.

Show counts: write “6 of 14” beside “43%,” every time.

Wait: look for the same direction across three cohorts before acting.

This is the book’s rule of thumb for small rooms. It doesn’t replace statistics. It keeps a busy owner from reading too much into a handful of people.

25.5 Using the Funnel to Decide

A funnel sheet is worth keeping because it tells you what to do next. After three or four cohorts, pick the stage with the largest loss in people and choose one change for that stage, using the tactics in Part Two. Write the change on the Chapter 4 plan, with the date, and mark the cohort that first meets it. Then keep counting. If the next cohort’s number moves by more than the normal swing, you’ve learned something. If it doesn’t, try another. Chapter 31 describes how to make that a fair test, so you aren’t fooled by a lucky month. Choose the change by the stage, using the funnel’s map of the book. A leak at the thirty-day return points to Chapter 10. A leak from returner to regular points to Chapters 9 and 7. A shortage of Ambassadors points to Chapter 7. Few first-timers point to Chapters 13 through 17.

Share the sheet. The host who greets, the teacher who runs the lesson, the bartender who sees the faces, and the booker who brings the acts should all see the numbers for their part of the funnel. People work better when they can see the result, and they often have the best ideas about the leak. Review it monthly in fifteen minutes, as Chapter 34 sets out. Don’t use it to blame. A low number is a prompt for a better question, and the people closest to the door usually know the answer before you do. When you share it, show the trend, not just the latest month. A teacher who sees that the second-lesson rate has risen three cohorts in a row can see her work paying off. A single month’s number, without the trend, can be discouraging even when things are improving.

Finally, tie the funnel to the money. Each regular is worth something over a year, and Chapter 32 shows how to work out what. If a change lifts the regular rate from 17% to 27% on cohorts of forty, that’s about four more regulars per month. Whatever a regular is worth, that’s a number you can compare with the cost of the change. The funnel turns what felt like a feeling into a decision, and a decision into a return. Everything else in Part Three, lists, content, press, partnerships, feeds into it, because each is a way of adding people to the top or keeping them from leaking out. The funnel sheet is also the best evidence you can show a partner, a sponsor, or a lender that the room is healthy, because it shows people choosing to come back, which no ad can buy. When you show it to someone outside the room, explain the regular rule and the counting method in one sentence at the top, so they read the numbers the way you do.

CASE STUDY

Three Cohorts at a Dancehall

Silas runs the Wednesday beginner night at a dancehall and hands every first-timer a cohort card. The January cohort has 40 first-timers. By the end of March he counts the returns by stamped cards: 12 within thirty days, 16 within sixty, 18 within ninety. 7 had three or more visits in sixty days, and 3 brought a friend who came. In February he adds a ninety-second welcome and a named first win. The February cohort has 44 first-timers: 19 back in thirty days, 24 in sixty, 26 in ninety, 12 regulars, and 5 ambassadors. March has 38 first-timers: 15 in thirty days, 20 in sixty, and 10 regulars so far.

Step 1 — January rates. 12 ÷ 40 = 30%. 16 ÷ 40 = 40%. 18 ÷ 40 = 45%. Regulars: 7 ÷ 40 = 17.5%. Ambassadors: 3 ÷ 40 = 7.5% of the cohort, or 3 ÷ 7 = 43% of regulars.

Step 2 — February rates. 19 ÷ 44 = 43.2%. 24 ÷ 44 = 54.5%. 26 ÷ 44 = 59.1%. Regulars: 12 ÷ 44 = 27.3%. Ambassadors: 5 ÷ 12 = 42% of regulars.

Step 3 — March, as far as it goes. 15 ÷ 38 = 39.5%. 20 ÷ 38 = 52.6%. Regulars: 10 ÷ 38 = 26.3%. Ninety days are not yet known.

Step 4 — The change. The thirty-day rate rose from 30% to 43%, a gain of 13 points. Regulars rose from 17.5% to 27.3%, about 10 points, or 12 − 7 = 5 more people on slightly larger cohorts. Visit frequency: January’s 18 returners made 41 more visits in ninety days, 41 ÷ 18 = 2.3 each. March held at 39.5% for thirty days, a little below February, which is inside the normal swing of his cohorts so far.

Step 5 — What he can and can’t say. The welcome and the first win were introduced together, and February had no holiday while January included a weekend rodeo. Three cohorts are a short series. He keeps counting and, next, tests the welcome alone, following Chapter 31. All figures are illustrative.

REAL TALK

From the dance floor:

I taught for years and thought I knew who came back. When I started stamping cards, I was wrong about nearly half of them. The ones I thought were shy had come four times. The ones I thought were sure things had vanished after the first night. The card was more honest than my memory.

HEADS UP

Following people over time is a form of data collection. Cards with initials are low risk, but sign-in sheets, loyalty features, and any record that ties a visit history to a name, phone number, or email trigger privacy and consent duties, including the Texas Data Privacy and Security Act for businesses it covers, federal CAN-SPAM for marketing email, and federal and Texas rules for marketing texts. Collect only what you need, keep it secure, and let people opt out. This is general information, not legal advice. Ask a licensed Texas attorney before building any record that identifies guests, and see Chapter 35 and Appendix J.

COHORT TRACKING SHEET

One sheet per night. Write your definitions first. A cohort is everyone who first came in the same period. Fill the empty cells as the clock runs out, and leave the rest blank.

Fill in: Night, Cohort period (week or month), Regular means.

Columns to fill in: Cohort, First-timers, Back in 30 days, Back in 60 days, Back in 90 days, Regulars, Ambassadors. Copy as many rows as you need.

Rate = count ÷ first-timers for that cohort. Returns are cumulative. Compare the same time point across cohorts, and don’t compare a young cohort’s 90-day number with an old one’s.

MEASURE IT

The number: The cohort sheet: for each month’s first-timers, the 30-, 60-, and 90-day return, the regulars count, visit frequency, and ambassadors.

By hand: Hand a dated card to every first-timer on your key night, and stamp it at each return. At the end of each month, count the first-timers, and then at 30, 60, and 90 days count the stamped cards that came back. Divide each by the cohort size. Example: 12 ÷ 40 = 30%. Count regulars (three visits in 60 days) and ambassadors (a regular who brought a guest who came).

Watch out: Cards get lost, so rates run a little low, consistently. Don’t compare a young cohort’s 90-day rate with an old one’s. Small cohorts swing, so look at people as well as percentages, and collect no more personal data than you need.

Go deeper: Chapter 31 (running experiments) and Chapter 32 (the money), plus Chapter 3 for the funnel stages.

ROLE BOX

• Owner: Pick your key night, hand out dated cards, and chart three cohorts before changing anything. Fix the biggest leak in people first.
• Booker: Run a cohort for acts: first-time acts each quarter, and how many were rebooked within 90 days and became regular dates.
• Musician: Run a cohort for rooms and fans: first dates in a room each quarter, and how many led to a second date. Track fans who come twice.
• Bartender: Treat guests who say they came because of your content as a cohort. Track how many return, and how many bring a friend.
• Dance Teacher: Your cohort is each month’s first lesson. Track the second lesson, the third in sixty days, and the step to the live floor.

TALK ABOUT IT

• If you had to guess, how many of last month’s first-timers have come back? How would you check?
• Which stage of your funnel loses the most people? Is it the one you spend the most effort on?
• What is the lightest way to count returners in your room that you’d still keep up for a year?