Attribution Without a Tech Stack

How’d you hear about us, promo codes, QR codes, tracked links, and the honest limits of each.

Every owner has said it at some point: “I know half my advertising works. I just don’t know which half.” The line is old, and it points at a real problem. A guest walks in on a Friday. Three weeks earlier a friend mentioned the place. Last week they saw a post. This morning they searched the name and found the map. Which one brought them? In a sense all of them did, and in another sense none can be proven. That puzzle is called attribution. Big companies spend a great deal of money on software that tries to answer it, and even they argue about the results. A small room can’t afford that software and doesn’t need it.

Big companies throw software at it. A small room can’t and doesn’t need to. What a small room can do is ask, count, and cross-check. The simple methods in this chapter, a question at the door, a code, a QR scan, a tagged link, a tally at the bar, won’t give you a perfect answer, but they’ll give you a useful one, and a useful guess beats a confident feeling. The key is to be honest about what each method can and can’t tell you. This chapter shows several simple methods: asking at the door, promo codes, QR codes, tracked links, and the data you may already have in a ticketing site or a register.

We’ll go through the five tools, show where each one fails, and then say what can’t be measured directly, because that list matters as much as the one of things that can. At the end you’ll have a “How’d You Hear” tally card, and Appendix E carries a printable one. The goal isn’t to credit each guest to exactly one cause. It’s to learn which of your efforts show up in the stories people tell about how they came. Along the way it shows where each method fails, how to cross-check one with another, and how to read the results without claiming more than they can support.

PR TACTIC: The Two-Second Question

What it is: Ask each first-timer, in the same words every time, “What made you decide to come tonight?” and tally the answer.

Why it works: Asking costs nothing, needs no software, and gets the answer in the guest’s own words, which is the closest thing to the truth you can collect.

Try this: Pick the exact wording tonight, write it on the tally card, and have everyone at the door and bar use the same sentence.

24.1 What Attribution Is, and What It Isn’t

Attribution means connecting a result to the thing that caused it. In a honky tonk the result is a person walking in, and the candidates for cause are everything they saw, heard, and were told before. Marketers talk about touches, each contact a person has with you before they act, and about first-touch and last-touch credit: the first thing they noticed, or the last thing before they came. Both are simplifications. Most guests have several touches, and the one they remember isn’t always the one that mattered. A friend’s recommendation may have worked because the person then saw the room looked busy online. That is why attribution in this book is about patterns, not verdicts. One guest saying a post brought them is a story. Twelve guests in a month saying it, compared with two the month before the posts began, is evidence worth acting on, even though no single answer can be proven.

So a small room’s attribution is always an estimate. Treat it as a way of ranking efforts, not of auditing them. If thirty percent of first-timers mention a friend and three percent mention a flyer, that’s a strong hint about where to put your effort, even if some of the “friend” guests also saw a flyer. The numbers won’t add up to a neat total, and they don’t need to. What you’re looking for is a pattern across weeks, and for changes in the pattern when you change something. It also helps to separate the question “did this effort bring anyone?” from the question “which effort deserves the most credit?” The first can often be answered with a modest count. The second rarely can, and pretending otherwise leads to arguments that nobody can win.

The biggest mistake is to confuse activity with arrival. Scans, clicks, likes, and views are things people did on a screen. A first-timer at the door is a person who came. The two are connected but not the same, and the gap between them is the subject of the rest of this chapter. A QR code scanned two hundred times that brought four people in is a four-person result. Always trace a number down to a person standing in your room, and be suspicious of any number that stops short of that. The same rule applies to the other chairs. A booker’s result is a confirmed date, not an opened email. A teacher’s result is a student in class, not a click on the class page. A musician’s result is a fan at the second show, not a stream. If a number can’t be traced to a person standing in the room, treat it as a clue about interest, not as a result.

24.2 Ask Them: The Door, the Bar, and the Tally

The simplest method is the best one: ask. At the door, the host or door person asks each first-timer one question, such as “What made you decide to come tonight?” and ticks a box on a card. The bar can do the same with new faces, and so can the teacher at the start of a class. Keep the wording identical, so the answers can be compared. Offer a short list if you like: a friend, a post, a search or map, drove by, a flyer or poster, news or radio, other. Let people name more than one. Write down “don’t remember” when that’s the answer, because it’s information too. Train the question so it sounds like hospitality, not a survey. “Welcome! First time here? How’d you hear about us?” said with a smile at the door is a greeting. The same words read off a clipboard are a form. Practice it once at the pre-shift huddle so the door person is comfortable.

Self-report has known flaws. People forget, and they tend to name the most recent or the most vivid thing. They also give polite answers: someone who found you on a map may say “a friend” because it feels warmer. And a busy door will produce ticks only for the people the host had time to ask. None of these ruins the method, and all of them argue for consistency. Ask the same way, to the same kind of guest, on the same nights, for a month or two. Track the share who answer, as well as the answers, so you know how much of the crowd the tally covers. A tally that covers one guest in ten isn’t a tally. It also helps to ask a follow-up when the answer is vague. If a guest says “online,” a gentle “Was it a friend’s post or one of ours?” often turns a useless answer into a useful one. Don’t press if they don’t remember. A response rate below about half of first-timers is a sign the question isn’t being asked consistently. Fix the asking before you read too much into the answers.

The tally also helps the staff. A host who asks every first-timer learns names, notices faces, and gets a chance to say something kind. It’s a welcome in disguise. Keep the card small and the question friendly, and don’t turn it into an interrogation. If you want to ask a second question, make it “Who did you come with?” which tells you whether guests arrive alone or in groups. Over a few weeks, the answers tell you which of your efforts people mention, which they never do, and what to ask the next group of guests. Appendix E has a card you can print. Ask staff to add any source they hear about during the night, such as a guest at the bar who mentions the radio interview. Those notes catch answers the door misses.

QUESTIONS PEOPLE ASK

“Won’t guests find the question annoying?” Rarely, if it’s asked warmly as part of a welcome. Most people like being asked how they found a place. Keep it to one question.

“What about regulars who bring a first-timer?” Ask the first-timer, and note the regular’s first name if the guest offers it. That’s how you find your Ambassadors.

24.3 Codes, QR Codes, and Tracked Links

A promo code is a word that guests say or show to claim an offer, and it works as a tag on the source. If the poster says “say BOOTS for a free beginner lesson” and the radio mention says “say HEARTLAND,” then the number of people who say each is a count of who responded to what. The code has to be easy to say and easy to track: the door person ticks a box. The trap is that a code measures only the people who use it. Guests who saw the poster but didn’t bother with the code are invisible, so the count is a floor. Also, an offer attached to alcohol raises rules; see the Heads Up below. Use codes that are easy to say and hard to confuse, and give each effort its own: one for the poster at the feed store, one for the radio mention, one for the partner’s newsletter. Keep a list of which code belongs to which effort, with the dates it ran, so the tally can be read months later.

A QR code is a square that a phone camera turns into a web address. Put a different code on each poster, table tent, or flyer, each pointing to the same page with a different tag, and you can see how many scans each got. That’s a measure of curiosity, not of arrival. A scan doesn’t mean a visit. To get closer, point the code at something that requires a step toward coming, like a sign-up for the Thursday lesson or a page with a code to say at the door. Then count the codes said, not the scans. Static codes that point directly to your own page are safer than codes that run through a generator’s service, which may stop working if a trial ends. Test every code from a phone before printing, and print the web address under the code for people who prefer to type.

A tracked link is an ordinary web address with a tag added to the end, which lets your analytics tool say where the click came from. These are often called UTM tags. If you share a link in your email, your social profile, and a partner’s page, each with its own tag, you can see which sent the most visits to your site. Free tools can build them, and most analytics tools read them. But tracked links only measure people who click and land on a page you control, and they miss everything that happens off the web. Treat them as one clue among several. A link click is a web event. A guest is a person. Keep a simple list of every tag you’ve used and what it meant, and use the same spelling every time. “Poster” and “poster-feedstore” and “Poster_Feed” will show up as three different sources in a report. Remove the tags from any link you share in a personal message, since they can look odd to the recipient and add nothing there.

24.4 Where Each Method Fails, and What You Can’t Measure

Each tool has a characteristic failure. The question at the door fails when the door is busy or the host forgets, and when guests give polite or vague answers. Promo codes fail when people forget to say them, when they’re too complicated, or when the code gets shared beyond its audience. QR codes fail when the scan is mistaken for a visit, and when the code is placed where nobody can use it, like on a billboard at highway speed. Tracked links fail when people copy the address instead of clicking, or when the platform hides where traffic came from. No method is reliable alone. Together, they cross-check one another. Ticketing and reservation sites fail differently. They can tell you who bought, when, and sometimes from which link, but not who actually came, who came without a ticket, or who bought for a group. Their reports also follow the platform’s definitions, which may change. Treat them as another clue, and check them against the door.

The cross-check is the best idea in this chapter. Use two methods for the same effort and compare. If a poster has a QR code and a promo code, and the door tally says six people came because of a poster, the code was said by four, and the scans were two hundred, then you can say something useful. Six people mentioned it, four proved it with a code, and most scans were curiosity. The overlap between self-report and a code is your most reliable evidence. A mention without a code is probably true, and a scan without a mention is probably noise. Over a few months, the cross-check also tells you how much to trust each method on its own. If the door tally consistently catches about half of the people who use a code, you can roughly adjust future tallies for efforts without codes, and say so when you report.

Be candid about what can’t be measured directly. A friend’s text message, a conversation in a truck, a recommendation at church, or a drive past a lit-up sign leave no trail. Marketers call the invisible part “dark” sharing. A great deal of word of mouth is dark, and the only way to see it is to ask. Don’t conclude that an untracked source doesn’t matter, and don’t invent a number for it. Report it as “unknown” and watch the size of the unknown over time. If it grows, you’re losing track of how people find you, which is a reason to ask more often. Some rooms add a line to the tally for this: “a friend told me, but I also saw it online.” It’s messy, and it’s true. Messy and true is better than tidy and false.

USING YOUR TICKETING AND REGISTER DATA

Ticketing or reservation site. Most offer a sales report by date. Some also show where buyers came from or let you make separate links or discount codes for each effort. Use those to compare efforts, and compare tickets sold with the door count to see how many people came without a ticket or didn’t show.

Point-of-sale register. Most can report sales and transactions by hour and by day, and some let staff press a button or add a note for a promo code. Use the transaction count as a check on the door count, and sales per head as a check on whether a promotion brought spenders or only bodies.

What they can’t tell you. Neither knows why someone came, whether they’d been before, or whether they’ll be back. Pair them with the how’d-you-hear question and the cohort card.

What to protect. Both hold personal and payment information. Export only the totals you need, don’t copy card or contact details into your own sheets, and see Chapter 35 and Appendix J.

TRY THIS WEEK

Print twenty how’d-you-hear cards. Hand one to every first-timer this Friday. On Saturday, count the answers and the share who answered. Don’t change anything else yet. Next week, do it again.

24.5 Reading the Tally: Rules of Thumb

Read the tally as shares, not counts. If sixty first-timers answered in a month and twenty-one named a friend, the friend share is 21 ÷ 60, which is 35%. Do the same for each source. Because guests can name more than one source, shares may add to more than 100%, which is fine. Compare month to month and look for a change bigger than the normal swing, as Chapter 23 explained. A source that goes from 10% to 12% is nothing. One that goes from 10% to 25% after you changed something is a result. Shares also need a stable base. If the share of first-timers who answered changes a lot from month to month, the shares will move for that reason alone. Report the response rate alongside the shares, such as “60 of 80 first-timers answered (75%),” so that a change in how hard the door asked isn’t mistaken for a change in how people found you.

Then ask what you’d do differently. If friends are the largest source, put effort into the Ambassador work from Chapter 7. If a post is cited more than anything else, make more of that kind of post. If “drove by” is large, check your sign and your outdoor lighting. If “don’t remember” is large, ask better or ask earlier. If one source you pay for is nearly absent, consider ending it, using the keep, fix, or kill review in Chapter 34. The tally is only useful if it prompts a decision. When two sources rise together, such as press and search, the second is often caused by the first. People read a story, then search your name. Count them separately, but read them together, and give the press some of the credit for the search.

Finally, keep your claims modest. “Thirty-five percent of first-timers who answered mentioned a friend” is accurate. “Thirty-five percent of our growth comes from word of mouth” is a leap. The tally shows what guests said, not what caused them to come. State it that way, and keep the number of people who answered beside it. A reader who sees “21 of 60” will judge the evidence fairly. A reader who sees only a percentage can’t. Chapter 25 uses the first-timer count to build the funnel, and Chapter 29 uses the same method for partnerships and promotions. Keep the tally sheets. A year of monthly tallies, side by side, is one of the most useful records a room can own, because it shows which sources grow, which fade, and which never mattered.

COMMON MISTAKES WITH ATTRIBUTION

Crediting the last touch. The thing a guest mentions is often the last thing they saw, not the first thing that made them curious. Read the answers as a mix, not a single cause.

Counting scans as guests. A scan is a phone pointed at a poster. A guest is a person at the door.

Dropping “don’t remember.” It’s an answer, and its size matters. If it grows, ask more carefully.

Changing the answer choices. New choices each month make months impossible to compare.

Claiming growth from shares. A source’s share can rise while its count falls, if the total fell faster. Look at both.

CASE STUDY

Sixty First-Timers and a QR Code

Yancey runs the door at a dancehall. For four weeks he asks every first-timer one question and ticks the answer. The room also put a QR code on its poster and a promo code on its radio mention. Of 72 first-timers, 60 answer. The answers: friend 21, social post 14, map or search 8, drove by 5, poster 3, radio 2, “don’t remember” 7.

Step 1 — Coverage. 60 ÷ 72 = 83% answered. Good enough to read.

Step 2 — Shares. Friend: 21 ÷ 60 = 35%. Post: 14 ÷ 60 = 23%. Map or search: 8 ÷ 60 = 13%. Drove by: 5 ÷ 60 = 8%. Poster: 3 ÷ 60 = 5%. Radio: 2 ÷ 60 = 3%. Don’t remember: 7 ÷ 60 = 12%. Check: 21 + 14 + 8 + 5 + 3 + 2 + 7 = 60.

Step 3 — The QR code. The poster’s code was scanned 212 times in the month. Only 3 first-timers named the poster. 3 ÷ 212 = 1.4%. The scans were mostly curiosity, many from people who had already come. The radio code was said by 2 people, and both also named radio at the door, so that source is confirmed.

Step 4 — The decision. Friends and posts together are 35% + 23% = 58% of answers, so Yancey puts effort into the Ambassador ladder and into clips, and stops printing posters for six weeks to see whether the poster share stays at 5%. He also reports the 12% “don’t remember” as unknown. A month of one night is a small sample, and he doesn’t treat the poster as useless yet. All figures are illustrative.

REAL TALK

From the booking desk:

I put a code on every poster for a year and proudly counted scans. Then the door person started asking first-timers how they’d heard. The scans said posters. The guests said their friends. I trusted the guests.

HEADS UP

Offers tied to codes can raise legal questions. Texas Alcoholic Beverage Commission rules limit on-premises alcohol promotions (see 16 Texas Administrative Code § 45.103), including free-drink and discount limits, so a code that unlocks a drink deal needs a check first. Tracking links and QR pages that collect names, emails, or phone numbers trigger consent and privacy duties, including federal CAN-SPAM rules for marketing email and federal and Texas rules for marketing texts. This is general information, not legal advice. Ask a licensed Texas attorney before pairing any code with alcohol or collecting contact details, and see Chapters 26 and 35 and Appendix J.

HOW’D YOU HEAR TALLY CARD

Ask every first-timer the same question and tick one or more boxes. Write the wording here so everyone uses it. Print more copies from Appendix E.

Fill in: Question wording, Night, Month.

• A friend or family member.
• A social post or video.
• Map or search.
• Drove or walked by.
• Poster or flyer.
• News, radio, or press.
• Other (write it).
• Don’t remember.
• First-timers asked / total.

Share = the source’s total ÷ first-timers who answered. Guests may name more than one source, so shares can add to more than 100%.

MEASURE IT

The number: Your source shares: for each source, the percentage of answering first-timers who named it, plus the share of first-timers who answered at all.

By hand: Tally one question for a month. Divide each source’s count by the number who answered. Divide the number who answered by all first-timers. Example: 21 ÷ 60 = 35% friend; 60 ÷ 72 = 83% answered. For any code or QR, count people who said it at the door, not scans.

Watch out: Scans, clicks, and views are not visits. Report “don’t remember” as unknown instead of spreading it across other sources. Never claim a share of growth the tally doesn’t show.

Go deeper: Chapter 25 (measuring the funnel) and Appendix E (the “How’d You Hear” card).

ROLE BOX

• Owner: Teach the door and bar one wording of the question, keep the tally card at the register, and review shares monthly.
• Booker: Ask each new act or partner how they heard of the room, and each first-timer at the door on your nights.
• Musician: Ask new fans how they found you at merch, and keep a code on your own links so a promoter’s numbers match yours.
• Bartender: Tally guests who mention your content and ask them what they saw. Count codes said, not views.
• Dance Teacher: Ask new students on sign-in how they heard. A friend’s invitation is the number to protect.

TALK ABOUT IT

• If you asked ten first-timers tonight how they heard about you, what do you think they’d say? What would surprise you?
• Which of your efforts produces numbers on a screen but no one at the door?
• How much of your crowd comes through channels you can’t see? What would you do if it were half?