Why Most Small-Business PR Measurement Fails

The numbers most rooms watch can’t tell them whether anything worked.

It’s a little after ten on a Friday, the band is on its second set, and the owner of a Central Texas dancehall is standing at the end of the bar with her phone in one hand and the door clicker in the other. The phone says the page picked up 40 new followers this month. The clicker says 212. The room feels good. And if you asked her, right there, whether the new Thursday promotion is working, whether the posts are worth the hours, whether the headliner is worth booking again, she would answer the way almost every owner, booker, musician, bartender, and dance teacher answers: she’d say she thinks so. She would be telling you a feeling, and the two numbers in her hands couldn’t confirm it or deny it.

None of this is a failure of the first five books. They pointed the right way. Volume One said likes and follower counts are pleasant but that people in the room are the goal, and told you to keep a loose eye on whether your messages put bodies through the door. Volume Two said to stop counting heads and start counting returns. Volume Three called the follower count the single most overrated number in a musician’s life, and noted that a band may want someone holding a clicker at the door so the count stays honest. Volume Four told bartender-creators to ignore the likes and read two numbers on every clip: how many people got past the first few seconds, and how many watched to the end. Volume Five built the Floor Funnel, four stages from First-Timer to Ambassador, and never put a number on any of them. Those were the right instincts. What none of them gave you was a method: a way to set a starting point, credit a result to the thing that caused it, test one change at a time, and decide to keep, fix, or drop a tactic. That is what this volume is for, and Part Three is the heart of it.

Before you can measure well, you have to see why people measure badly, and that’s this chapter. There are five ways small-business PR measurement goes wrong, and nearly every honky tonk, booking calendar, band, bar, and dance floor commits at least three of them: chasing numbers that feel good instead of numbers that decide something; reading a result with no starting point to compare it to; watching so many numbers that something always looks like a win; mistaking ordinary wobble for a result; and turning a number into a target until it stops telling the truth. We take them in that order, and the chapter ends with a five-question test you can run on any tactic in this series. A word on the math: every calculation in this book is adding, subtracting, multiplying, dividing, and percentages — the kind you can do on a bar napkin or in a spreadsheet — and every worked example is an illustrative composite with the arithmetic shown. Any outside fact is sourced in the back of the book.

PR TACTIC: The Written Win

What it is: Before you try anything, write down the one result you want, the number that would show it, and what you will do if it doesn’t happen.

Why it works: A target set in advance can’t be bent to fit whatever happens, and writing it forces you to name the goal a tactic is supposed to serve.

Try this: Put one sentence on an index card before your next promotion: “If this works, [number] will be at least [amount] by [date]; if it doesn’t, we will [decision].”

21.1 The Number That Feels Good

A vanity metric is a number that looks like success without being able to tell you whether you’ve had any. Eric Ries, the author of The Lean Startup, put the warning in three words: vanity metrics are dangerous. They’re dangerous for a reason that has nothing to do with anyone lying. Followers, likes, impressions, total reach, and the all-time count of people who have walked through your door share three traits. They’re easy to get, they’re always sitting on a screen or a clicker where you can see them, and they almost never go down. A number that only goes up will make you feel good every time you look at it, and a number you feel good about stops being questioned. That is the trap. You will never be fooled by a number that disappoints you. You will be fooled by the one that flatters you.

Here’s a ten-second test that separates a vanity number from a useful one: could this number go up while the room gets worse? If the answer is yes, it’s a vanity number. Followers can climb while nobody comes. Reach can climb while nobody reads. And the head count — the most honest-feeling number a room owns — can climb while the room empties, which is what the showcase owner in Volume Two discovered when five acts’ friends took turns coming through the door. The clicker said the night was a success; the bar said otherwise. A head count counts visits, not people. It can’t tell a hundred different guests from fifty guests who each walked in twice, and it can’t tell a first-timer from a regular you’ve known for six years. Counting returns, the fix Volume Two urged, solves exactly that, and Chapter 25 shows you how to do it by hand.

Even a number that looks useful can quietly go bad underneath you. Take the email open rate, which Volume One listed among the signals worth a glance. Apple’s Mail Privacy Protection, which arrived with iOS 15 in September 2021, is described by Apple as helping “users prevent senders from knowing when they open an email,” and email services such as Mailchimp warn that marketers may see inflated open rates because of it. In plain terms: for a reader using that setting, your email tool can record an “open” that no human performed. So the open rate, once a rough stand-in for “people are paying attention,” now partly counts machines. Clicks, and better still people showing up, are stronger evidence, and Chapter 26 goes through the details. The lesson isn’t to memorize one company’s settings. It’s to ask, every time, what a number actually counts, and who or what is doing the counting.

REAL TALK

From behind the bar:

I had a clip hit 40,000 views once. Forty thousand! I walked around like I’d bought the place. The next Friday I asked every table I could, “You see me online?” — and in a whole night exactly two people had. One of them was my cousin. The clip that did the real work was a thirty-second one about how I clear a rail at close, which got maybe nine hundred views, and I know it worked because three different regulars brought it up by name. So I started a tally on a bar napkin: “Saw you online” on one side, “Just here” on the other. That napkin taught me more than the whole dashboard ever did.

21.2 No Baseline, No Answer

A number by itself answers nothing. “We had 140 on Friday” is neither good news nor bad news until you know what a Friday usually brings. A baseline is simply the normal level of the thing you’re measuring, worked out before you change anything, and it is the most skipped step in small-business PR. Without one, you don’t measure a promotion; you judge it by mood. Mood is generous to whatever you just paid for, because nobody wants to feel that the $150 was wasted, and it’s hard on whatever you didn’t try. Chapter 23 builds a proper baseline, and Appendix B gives you the worksheet. Here we only need the idea: you can’t see a change until you’ve seen what it changed from.

A baseline has two parts, and the second is the one almost nobody learns. The first is the typical level, which you get by adding up several comparable nights and dividing by the number of nights. The second is the normal swing, which is simply the lowest and the highest those nights reached. Say the last six Thursdays at a small dancehall drew 58, 64, 71, 55, 69, and 61 people. Add them: 58 + 64 = 122, plus 71 is 193, plus 55 is 248, plus 69 is 317, plus 61 is 378. Divide 378 by 6 and the typical Thursday is 63. The swing runs from 55 to 71. That tells you a Thursday of 66 is not news, a Thursday of 58 is not a disaster, and a Thursday of 75 is the first number in the series that deserves a real question.

The comparison has to be fair, and that’s where baselines get bent. Compare a Thursday only to other Thursdays, never to a Saturday, and never to the Thursday of a rodeo weekend, a holiday, a payday week, a big Friday-night football game, or the night the creek came up and half the county stayed home — not without marking it. Compare a music night to a music night of the same kind, and a beginner lesson to a beginner lesson. Every owner and bartender already knows this in their bones; you can tell by five o’clock how the night will run. The point of writing it down is to stop leaving that knowledge in your head, where it only gets used when it protects your pride. “Slow because of the weather” is a fair note when you wrote down the weather before you knew the result, and a convenient excuse when you think of it afterward.

21.3 Too Many Numbers, No Goal

The third way to lose the thread is to measure everything. Every platform you use hands you a dashboard before you’ve asked it a question: reach, impressions, profile visits, link clicks, shares, saves, follower change, top posts, best times to post. Your point-of-sale system adds sales by the hour, by the item, by the server. Your email tool adds opens, clicks, and unsubscribes. None of it is wrong, and most of it is free, which is exactly the problem. A dashboard is built around what the platform can count, not around what you’re trying to do, and the number it puts at the top of the screen is the number the platform wants you to care about. When you don’t have a goal of your own, you inherit theirs.

Here is the arithmetic of why that matters. Suppose you track twelve numbers each week, and, to keep the math simple, treat each one as a coin flip that has even odds of ticking up in any given week purely by chance. The chance that a single number goes down is 1 in 2. The chance that all twelve go down is 1 in 2 multiplied by itself twelve times, which is 1 in 4,096. So the chance that at least one number went up is 4,095 in 4,096. Real numbers aren’t perfectly independent coin flips, so the exact odds are fuzzier, but the lesson holds: if you go looking through twelve numbers for good news, you’ll find some every single week, whether or not anything you did mattered. People call this cherry-picking, and it’s seldom done on purpose. It’s just what happens when you look at the dashboard first and decide what you were trying to do second.

The fix is old-fashioned: decide the goal first, then pick the numbers. Chapter 4 gave you the one-page plan that runs from goal to audience to message to channel to measure, and Chapter 22 shows how to turn a goal into five numbers or fewer. For now, borrow one habit. Before you start any tactic from this series, write on a card the single result you want, the number that would show it, and what you’ll do if it doesn’t happen. That card is the PR Tactic at the top of this chapter. Everything else on the dashboard can be glanced at, enjoyed, and set aside. A number that didn’t make the card is context. It isn’t evidence.

21.4 Noise Mistaken for Results

Every room wobbles. Two Thursdays that look identical on paper — same band, same weather, same promotion — won’t draw the same crowd, and neither will two Tuesdays or two lessons. A cousin’s birthday, a flat tire on the highway, a better show across town, a bad day at work: a night’s attendance is the sum of a hundred small decisions made by people you’ll never meet. That ordinary wobble is called noise, and noise is the great counterfeiter of results. When a number moves inside its normal swing, nothing has been shown, however good it feels. A useful working rule — this book’s own rule of thumb, not a statistical law — is that a change smaller than the normal swing of the baseline is not yet a result.

Noise fools us in a particular way that’s worth naming. After an unusually bad night, the next one is likely to be better, because very bad nights are rare and ordinary nights are the rule. If you made a change after the bad night, the change gets the credit. Say a Wednesday that normally draws about 40 drops to 22 because of a storm. You rework the playlist, post a new flyer, and the next Wednesday 38 people show up. It looks like a rescue: 38 − 22 = 16 more people, which is 16 ÷ 22 = 73% more than the bad night. But measured against the usual 40, nothing happened. The reverse trap is just as common: a lucky night is credited to whatever you did that week, and the ordinary night after it makes the tactic look like it wore off. In both directions the cure is the same. Compare to the baseline, not to last week.

Two more habits manufacture false results. The first is changing several things at once. If the same week you launch a new Wednesday special you also get a write-up in the weekly paper, a band’s post goes around, and the weather finally breaks, you’ll see a bump and have no way to know which of the four earned it. Chapter 31 shows you how to change one thing at a time. The second is reading too much into small counts. Three more guests out of 40 is a 7.5% increase on paper and a coin flip in the room, and a dance teacher whose Tuesday class goes from 11 students to 13 has learned almost nothing. The smaller the room, the bigger the swing looks as a share of the total, which means small rooms and small classes need longer stretches of data, not fancier math. The honest answer to “did it work?” is often “we can’t tell yet,” and a measurement habit that can say that out loud is already ahead of most.

CASE STUDY

The Boosted Thursday

(A composite, drawn from a pattern common to Texas rooms and the people who run them; the people are illustrative, the lessons are real.)

Colette runs a mid-size dancehall in Central Texas. Thursday is her slowest night worth keeping, so she spends $150 boosting one Facebook post for the coming week. The post shows 18,400 people reached and 612 reactions. That Thursday the door clicker reads 94. She tells her bartender, Tobin, that it worked and that she’s going to double the budget. Tobin keeps a notebook behind the bar and asks if he can run the numbers first.

Step 1 — The baseline from the clicker. The previous eight Thursdays read 71, 83, 66, 90, 78, 69, 85, and 74. Add them: 71 + 83 = 154, plus 66 is 220, plus 90 is 310, plus 78 is 388, plus 69 is 457, plus 85 is 542, plus 74 is 616. Divide 616 by 8 and the typical Thursday is 77. The normal swing runs from 66 to 90.

Step 2 — The claim. 94 − 77 = 17 more than typical, and 17 ÷ 77 = 22%. That sounds large. But 94 is only 4 above the best Thursday in the baseline (90), so it barely clears the swing.

Step 3 — Check the clicker against the tabs. The clicker counts every click at the door, including guests who stepped out to the patio and came back in, so Tobin also pulls the number of tabs opened. That Thursday: 71 tabs. The previous eight Thursdays: 58, 66, 52, 70, 63, 55, 68, and 60. Add them: 58 + 66 = 124, plus 52 is 176, plus 70 is 246, plus 63 is 309, plus 55 is 364, plus 68 is 432, plus 60 is 492. Divide 492 by 8 and the typical Thursday is 61.5 tabs, with a swing from 52 to 70. So 71 − 61.5 = 9.5 more tabs, and 9.5 ÷ 61.5 = 15%. The clicker said 22%; the tabs say 15%, and 71 is just one above the highest tab count in the baseline.

Step 4 — Reach against results. 18,400 people reached ÷ 9.5 extra tabs is roughly one extra tab for every 1,900 people reached. And $150 ÷ 9.5 = about $15.79 for each extra tab. That isn’t automatically a bad price, but it’s a very different story from “612 reactions.”

Step 5 — What else changed. That Thursday was a payday, and the county fair opened up the road the same night. Either could have lifted the crowd with or without the post. Run through the five questions at the end of this chapter, and the promotion comes out not yet proven: no written goal beforehand, a result barely outside the swing, a likelier explanation sitting right there, and a clicker that overstates the gain. Tobin’s recommendation is the humble one. Run the same $150 post on an ordinary Thursday two weeks out, with the goal written down first, and count tabs. If the number lands near 71 again, Colette has something. If it lands back near 62, she saved $150 and learned a lesson that cost nothing.

21.5 When the Number Becomes the Target

The last failure is the sneakiest, because it happens after you’ve done everything else right. An economist named Charles Goodhart noticed back in 1975 that an observed statistical regularity tends to collapse once pressure is placed on it for control purposes. The plain-language version, usually credited to the anthropologist Marilyn Strathern, who wrote it in a 1997 paper on auditing British universities, is this: “When a measure becomes a target, it ceases to be a good measure.” A famous case is Wells Fargo. In 2016 the Consumer Financial Protection Bureau found that, “spurred by sales targets and compensation incentives,” employees opened more than two million deposit and credit card accounts that may not have been authorized by consumers. The sales numbers looked wonderful right up until the moment people checked what was behind them.

You don’t need a bank for this to happen in a honky tonk; a free-drink coupon and a clipboard will do it. Pay the bartenders a bonus for every “found you on Instagram” on the tally, and the tally fills up with Instagram. Judge a booker by Friday head counts, and the booker starts stacking five opening acts for their friends, or leaning on bands to sell tickets, which is the very thing the showcase owner in Volume Two admitted to before tearing the night down and rebuilding it. Offer a free drink for a follow and you’ll get the followers and none of the faces. Discount the slow Tuesday and attendance rises with regulars you’ve just taught to wait for a deal. And shave the door count to pay a band less, which Volume Three called one of the worst trades in the business, or pad it to look busier, and you’ve spent the one thing a band uses to decide whether to trust your room again.

You can’t escape targets, because a number nobody cares about changes nothing. What you can do is build guardrails. First, pair every number with a check on its quality: if you count new followers, also count how many of them turned into visits; if you count entries at the door, also count tabs opened, the way Tobin did. Second, keep the person who counts separate from the person who’s graded on the count when you can, and write down exactly how the count is made so it’s made the same way every time. Third, decide in advance what you’ll do when the result is bad, because a number you’ve promised yourself to act on honestly in both directions is much harder to bend. Chapter 35 returns to this under honest numbers. And that brings us to the tool that ties the chapter together: five questions to ask of any tactic before you spend another dollar on it.

HEADS UP

The moment you start asking guests how they heard about you, collecting emails, or texting phone numbers, you’re handling personal information, and rules apply. Email marketing falls under the federal CAN-SPAM Act, marketing text messages under the Telephone Consumer Protection Act and carrier rules, and Texas has its own law on protecting sensitive personal information. A door count that sets a band’s pay is also part of a contract, so write down how the count will be made in the deal memo (Chapter 12). This is general information, not legal advice. Chapters 26 and 35 and Appendix J go further; before you launch a text club or a sign-up form, have a Texas attorney review the wording.

THE FIVE-QUESTION “DID IT WORK” TEST

Run this on any tactic from Volumes 1–5 after it has had time to work. Write the answers down; don’t just think them.

Fill in: Tactic, Dates.

• 1. Goal. What exactly were we trying to change, and did we write it down before we started? (Yes / Not sure / No)
• 2. Baseline. What was the number before, and what is its normal swing (lowest to highest)? (Yes / Not sure / No)
• 3. Size. Did it move by more than the normal swing — and, unless it was a one-time event, on more than one night? (Yes / Not sure / No)
• 4. Cause. What else changed or happened — weather, holiday, payday, another event, a second change from us — that could explain it? (Yes / Not sure / No)
• 5. Honesty. Did we count it the same way we would have if it had gone badly, and does what staff and regulars saw in the room agree with the number? (Yes / Not sure / No)

Reading your answers

• Five clear yeses: keep it going and move it to the keep-fix-kill review in Chapter 34.
• Three or four: not yet proven. Run it once more and change nothing else.
• Two or fewer: you don’t know yet. The honest label is unproven, not failed.

This scoring is the book’s own rule of thumb for small rooms, not a statistical law.

MEASURE IT

The number: Your Number Audit score: of every number you look at in an ordinary week, the share that has a written goal, a known normal swing, and a decision attached to it.

By hand: List every number you glance at in one week — followers, door count, tabs, open rate, anything. Beside each, mark G if it has a written goal, S if you know its normal swing, and D if a high or low reading would change something you do. Count the numbers marked G, S, and D, then divide by the total. Example: 9 numbers watched, 2 with all three marks, so 2 ÷ 9 = 22%. Every number short of three marks is context, not evidence.

Watch out: Don’t chase a perfect score by deleting numbers you find useful. The goal is fewer, better numbers, not a bare dashboard.

Go deeper: Chapter 22 (pick the number that matters) and Chapter 23 (build the baseline).

ROLE BOX

• Owner: The tempting number is the door clicker. Add a tab count beside it every night and mark first-timers versus returners, so a busy night and a good night stop looking the same.
• Booker: The tempting number is one night’s door. Log each booking against that weekday’s baseline, and note how much of the crowd was new to the room, before you decide an act drew or flopped.
• Musician: The tempting number is followers and streams. Track what a booker can act on instead: pitches sent, replies, rebookings, and how many in your crowd were new to the room.
• Bartender: The tempting number is clip views. Keep the napkin tally of “saw you online” versus “just here,” and read completion and sends on your clips, not likes.
• Dance Teacher: The tempting number is how many came to the lesson. Count how many of them are on the floor next week, because returns are what fill a floor.

TALK ABOUT IT

• Pick the one number you check most often. Could it go up while your room, your calendar, or your class got worse? What would that look like?
• Do you know the normal swing of your slowest night, lowest to highest, over the last eight weeks? If not, what would it take to find out this week?
• Think of the last time you decided a tactic “worked.” Which of the five questions could you have answered then, and which would you have had to guess at?
• Is anyone on your team paid, praised, or pressured over a number? What might they do to make it look better, without meaning any harm?