Measuring Content and Social

Retention curves, sends and saves over likes, profile visits, and the gap between followers and people who show up.

Every platform gives you a dashboard, and every dashboard is built to keep you looking at it. Reach, impressions, views, likes, shares, saves, follows, profile visits, top posts, best times to post. The screen is full, and the numbers are mostly true. But true isn’t the same as useful. Chapter 22 told you to pick the numbers that lead to a decision. This chapter applies that to content: which few numbers on a platform’s analytics page tell you whether your posts are doing their job? The trouble is that the page is full. Every platform offers dozens of figures, and most of them describe how a post did on the platform, not whether it helped the room.

The job, as Chapter 14 argued, is to show people what the night is like and send some of them to the door. So the measures that matter are the ones closest to that: whether people watch to the end, whether they send a post to someone, whether they save it, whether they visit your profile, and, finally, whether they come. The rest, likes, impressions, and total followers, are context. They’re easy to count and easy to like. They’re weak at predicting a full room. The useful numbers are few: whether people watched, whether they passed it on, whether they took a step toward you, and, finally, whether they came.

We’ll also be frank about the limits. Platforms define their numbers in their own way and change them. They tell you what happened on the platform and nothing about the room. And a post that did well last week may do badly this week for reasons nobody outside the company can see. So you’ll learn to read your own analytics carefully, keep a simple content log, and compare yourself with yourself. The aim isn’t to master an algorithm. It’s to find out which of your posts bring people through the door, and to make more of those. We’ll go through what the platforms count and why it differs, the numbers that show a post traveled, the chain from a profile to the room, how to read your own analytics, and a monthly rhythm for deciding what to make.

PR TACTIC: The Content Log

What it is: One line per post: date, format, the one thing it was meant to do, reach, sends, saves, average watch time, and whether anyone mentioned it at the door.

Why it works: A log turns scattered posts into a record you can learn from, and it shows the pattern a single dashboard screen can’t.

Try this: Start a spreadsheet tonight with those columns, and fill in your last ten posts from the platform’s analytics.

27.1 What the Platforms Count, and Why It Differs

Each platform has its own words and its own definitions. Reach usually means the number of different accounts that saw a post, while impressions or views may count every time it was shown, including repeats. Views may count a play after one second or after three. Watch time, retention, and completion mean different things on different services. YouTube, Instagram, TikTok, and Facebook each publish help pages that define their metrics, and the definitions change. Before you compare two numbers, read what each platform says it counts. A number from one platform can’t be set beside a number from another unless the definitions match, and they usually don’t. Definitions also change. Platforms rename metrics, merge them, or change how they’re counted, often with little notice. Note the date when you start tracking a number, and when a figure suddenly jumps or drops across all your posts at once, check whether the platform changed its counting before you assume your audience changed.

Platforms also share some of what their systems value. In January 2025 Instagram’s head, Adam Mosseri, said in a video that the top ranking signals are watch time, likes, and sends, and he advised creators to watch average watch time, likes per reach, and sends per reach. That’s a statement about one platform at one moment, and it’s useful because it points at the metrics worth watching. It also fits a larger pattern: platforms reward content people finish and pass along. For your purposes, watch time and sends are the numbers closest to the idea that a post was worth someone’s attention. YouTube’s help pages, for example, describe average view duration as the average minutes watched among viewers who stayed to watch, and describe the audience retention report as showing how well different moments of a video held attention. Those definitions are useful because they’re specific, and they tell you exactly what a change in the number means.

Three cautions. First, analytics only cover people on the platform: they say nothing about what happens when someone puts the phone down and comes to your room. Second, they’re delayed and sometimes revised, so check a post a few days later. Third, they’re calculated on the platform’s terms, so treat the numbers as indicators, not as accounting. The most reliable evidence is still the door: what first-timers say, and what the funnel from Chapter 25 shows. Content measurement is about finding out which posts to make more of, and the door confirms whether it worked. Fourth, the numbers describe a sample of people who happened to see a post on a particular day, shaped by a system you can’t see. Treat them as signals to compare among your own posts, not as precise measurements. Platforms also report differently for paid and unpaid posts. Keep boosted posts on their own lines in the log, so their numbers don’t inflate your normal.

WHAT THE MAIN PLATFORMS REPORT, IN PLAIN WORDS

Definitions change, so read each platform’s current help pages. As of the sources checked for this book:

YouTube: average view duration is the average minutes watched among viewers who stayed to watch. The audience retention report shows how well each moment held attention. A flat line means people watched that part through, and a dip means they left or skipped. Impressions click-through rate measures how often people watched after seeing a counted impression on YouTube, and not every glimpse of a thumbnail counts as an impression.

Instagram: its head said in January 2025 that the top ranking signals are watch time, likes, and sends, and he suggested creators watch average watch time, likes per reach, and sends per reach.

TikTok: creator analytics include average watch time, the share who watched the full video, the number of unique users reached, total play time, and video views, according to a guide by the social media company Buffer. We couldn’t load TikTok’s own help page, so treat these as a secondary description and check TikTok’s current help pages.

Map listing: Google’s Business Profile performance report includes searches, views, direction requests, calls, website clicks, and messages, with views counted once per person per day.

27.2 Retention, Sends, and Saves: The Numbers That Travel

For video, the most useful picture is the retention curve: a line showing what share of viewers are still watching at each moment. Most platforms offer it in some form. A steep drop in the first seconds means the opening failed. A gentle slope that ends high means the clip held people. A bump or spike near the end may mean people replayed it. The curve tells you where a clip loses its audience, and it ties directly to the Clip Formula in Chapter 14: if viewers leave at the third second, rework the hook. You don’t need a number to read it, just the shape. On YouTube, a flat stretch of the retention line means viewers watched that part through, and a dip means they were leaving or skipping, as YouTube’s own help page explains. Other platforms show similar curves or a single figure, such as the share who watched to the end. Use whichever your platform gives you, and read it for the moment people leave.

The second is sends per reach: the number of times people sent a post to someone, divided by the number of people it reached. A send is a recommendation. It’s word of mouth in digital form, and Chapter 7 showed why that matters. Close behind is saves per reach, which suggests the post was useful enough to keep. Both are better than likes, because they take a deliberate action and carry a social meaning. Compare them across your own posts. If your room clips get a 1.5% send rate and your graphics get 0.5%, you have an answer about what to make. Saves are a cousin of sends. A save usually means someone wants to come back to a post, such as a lesson clip they want to practice or a calendar they want to remember. For a teacher, saves on teaching clips are a good sign of usefulness. For a room, saves on a calendar post suggest people are planning around it.

Likes are the weakest of the common signals, because they’re cheap, and a high like count often reflects an audience already fond of you rather than new people. Comments are mixed: a thoughtful comment shows interest, but comments can be provoked by controversy that brings the wrong crowd. Follower growth is the most misleading of all, because it measures a one-time decision, not whether the person ever comes. The question to ask of every number is “out of whom, and then what?” Sends per reach tells you what fraction of viewers acted. A like total tells you nothing about the viewers who didn’t. Comments are worth reading even when they’re few, because they tell you what people wonder about. A clip whose comments ask “where is this?” or “is this every week?” is doing its job and needs a clearer caption. Questions in comments also make good next posts. If three people ask whether beginners are welcome, the next clip should answer that in the first three seconds.

27.3 From Profile Visits to People in the Room

Between the post and the door lies a chain: seeing, watching, visiting your profile, following or clicking, and finally coming. Platforms show some of the early links. Profile visits show how many people looked at your page after seeing a post, a stronger sign of interest than a view. Link clicks show how many tapped through to your calendar or site. Follows show how many agreed to see more. Each link loses people, and the rates between them tell you where the chain is weak. A post with great reach and few profile visits has an interesting clip but no pull toward you. A page with many visits and few clicks has a profile that isn’t giving the next step. Each link can be measured roughly from what the platforms show. Profile visits divided by reach shows how many viewers were curious enough to look you up. Link clicks or follows divided by profile visits shows how many found a reason to stay in touch. Neither is precise, but both show where the chain breaks.

The final link is the one platforms can’t see: the door. This is where the first-timer question from Chapter 24 matters. If you gained 120 followers this month and, of 40 first-timers who answered at the door, 6 said they came because of a post or video, the follower-to-visitor gap is plain. A hundred and twenty people said yes to your content, and a handful acted. That’s not unusual, since most followers are in other towns or never intended to visit. But the gap tells you to measure the post-to-door rate, not followers, and to aim your content at people near enough to come. That gap is one of the most useful content numbers a room can have, because no platform will ever report it. Divide the number of first-timers who mentioned your content in a month by the number of followers you gained in the same month. It won’t be precise, but over several months it tells you whether your audience is turning into guests.

Local reach matters more than total reach. If your analytics show where viewers live, check the share within driving distance. A clip seen by a million people in other states may produce no guests, while one seen by three thousand people within thirty miles may fill a Thursday. Tailor the content and the call to action to the people who can walk in: say the town, the day, the time, and what to do. Use the Discovery Signal from Chapter 13. The best content measure for a local room is the number of guests who say a post brought them, which the funnel then carries forward. If your platform doesn’t show location, watch the comments and messages for clues, such as people asking how far you are from a nearby town, and ask at the door how far guests drove.

QUESTIONS PEOPLE ASK

“Should I pay to boost the post that did best?” Only if it’s already bringing people through the door, not just views. Chapter 2’s amplifier rule applies: pay behind what already works.

“My followers are mostly out of town. Is that bad?” Not bad, but not a source of guests. If the goal is the room, aim more posts at people within driving distance, with clear local details.

27.4 Reading Your Own Analytics Without Fooling Yourself

Posts vary a great deal, and one outlier can mislead. A single clip that takes off isn’t a trend, and a single flop isn’t a verdict. To read content fairly, compare a post against your own recent posts of the same kind, in the same format and on the same platform. The baseline idea from Chapter 23 applies directly: find the average sends per reach for your last eight room clips and the normal swing. A post inside the swing is ordinary. One far above it is worth studying, and one far below is worth a gentle look at the first three seconds. A simple way to find your normal is to list the last twenty posts of one kind, such as room clips, and their sends per reach, then find the average, the median, and the range, exactly as Chapter 23 did for door counts. That gives you a baseline to judge the next post against.

Be careful about timing, too. A post made the day before a big local event, or during a holiday, will behave differently from one made on an ordinary Tuesday. Note unusual weeks in the log, as you would in a baseline. And be careful about changing several things at once. If you change the hook, the length, and the posting time together, you can’t tell which mattered. The experiment method in Chapter 31 applies: change one thing, keep the rest, and compare. A month of such steady testing teaches more than a year of guessing. Separate kinds of posts when you compare. A teaching clip, a calendar graphic, a regular’s story, and a band clip are different products with different jobs. Comparing a calendar post’s sends with a funny clip’s sends tells you nothing useful. Comparing calendar posts with each other does.

Remember Goodhart’s law again. If you chase sends alone, you may drift toward content that’s shareable and off-brand, such as stunts that bring the wrong people. Pair every platform number with a door number. Sends with first-timers who mention a post. Followers with followers who visit. If the two move together, the content is working. If sends rise and the door stays flat, you’re building an audience that isn’t yours. In that case, go back to Chapter 14 and ask whether the content shows the room as it is and invites people toward it. Ask the house-alignment question from Chapter 5 about every post that performs unusually well: does this show the room as it really is, and would the owner be glad to see it? A post that travels for the wrong reasons brings the wrong guests.

COMMON MISTAKES WITH CONTENT NUMBERS

Comparing across platforms. Each counts views, reach, and engagement differently.

Chasing the outlier. One clip that took off is a story to study, not a strategy to copy blindly.

Ignoring location. Reach outside driving distance rarely becomes a guest.

Counting likes as success. Watch time, sends, and guests at the door say much more.

Forgetting the hours. A clip that brought two guests and took six hours may not be worth repeating.

Not asking at the door. No dashboard will ever tell you who came in.

27.5 A Monthly Rhythm: Log, Review, Decide

Set a rhythm you can keep. After each post, or once a week, fill in the log: reach, sends, saves, average watch time, and anything a guest said. Once a month, spend twenty minutes reading it. Sort the posts by sends per reach and look at the top three and the bottom three. What did the best have in common: the opening, the subject, the length, the people? What did the worst share? Write one sentence of what you learned and one thing you’ll do differently next month. A log kept this way is a small research project, and over a year it’s more reliable than any outside advice. Keep the review short and the same each month. Look at the log, find the two posts that did best and the two that did worst on sends per reach and on guests mentioned, and ask what made the difference. Write one sentence for each. Over a year those sentences become your own guide to content.

Use the findings for the Platform Triage decision in Chapter 14. After sixty days, ask of each platform whether it moved anything real: guests who mention it, replies, bookings. If the answer is no, cut it, test it, or keep it with a lower expectation. Don’t keep a platform because you’ve always been on it or because a competitor is. Time is the scarce resource, and a platform that brings applause but no guests costs more than it earns. A shorter list of platforms, done steadily and measured honestly, will beat a scattered presence nearly every time. When you drop a platform, say so on it, pin a post that says where to find you instead, and keep the account so nobody else can take the name.

Finally, keep a record of the top posts for the proof file in Chapter 6 and for the story bank in Chapter 8. A clip that produced sends and guests is evidence. It’s also the raw material for the next pitch, bio, or one-sheet. Content measurement isn’t the end of the process, it’s the start of the next piece of work. Fill in the content log below, run it for a month, and see what your own numbers say about what to make next. Chapter 34 shows how to pull the few numbers that matter onto the monthly scorecard. Content work is also time work. Track the hours in the same log, and divide guests mentioned by hours spent each month. That figure, guests per content hour, is the one that tells a bartender or a teacher whether the work is worth continuing at its current pace. Share the month’s best post and what it taught you with the owner and the staff. Good content ideas often come from the people who saw the moment happen.

TRY THIS WEEK

List your last ten posts in a notebook with reach, sends, and average watch time. Mark the two best by sends per reach. Ask at the door this week whether anyone saw them. Write down what the two have in common.

CASE STUDY

Twelve Posts, One Pattern

Mabel runs a bartender’s page for a honky tonk. She logs twelve posts over a month: six room clips and six graphics. The six clips reached 5,400 people and were sent 81 times and saved 54 times. The six graphics reached 3,600 people and were sent 18 times and saved 14 times. She also gained 90 followers. Of 32 first-timers who answered at the door that month, 8 named a post or video.

Step 1 — Sends per reach. Clips: 81 ÷ 5,400 = 1.5%. Graphics: 18 ÷ 3,600 = 0.5%. The clips are sent three times as often per person reached.

Step 2 — Saves per reach. Clips: 54 ÷ 5,400 = 1.0%. Graphics: 14 ÷ 3,600 = 0.4%.

Step 3 — The door. 8 of 32 first-timers = 25% named a post or video. She asks which kind: 6 named a clip and 2 a graphic, so clips account for 6 ÷ 8 = 75% of the post-driven guests.

Step 4 — The gap. 90 new followers, and 8 first-timers from posts. 8 ÷ 90 = about 9%, though not all 8 were new followers. It’s a rough gap, not a rate. She decides to make clips two thirds of her posts for the next quarter and to log the next month the same way. A month of twelve posts is a small sample, and some clips were about a larger night. All figures are illustrative.

REAL TALK

From behind the bar:

I was proud of my follower count until I asked who’d actually been in. Out of forty guests, six said they’d seen me online. The rest were friends and walk-ins. I stopped counting followers that day and started counting faces.

HEADS UP

Platform analytics can include data about identifiable people, and what you do with it is regulated. Retargeting, custom audiences, and uploading a guest list to an ad platform raise privacy and consent questions, including under the Texas Data Privacy and Security Act for businesses it covers. Paid promotion and endorsements must be disclosed under Federal Trade Commission rules, and the FTC’s rule on reviews and testimonials bans buying fake followers or views. Alcohol-related content may also face platform and TABC restrictions. This is general information, not legal advice. Ask a licensed Texas attorney before uploading guest data to any platform, and see Chapter 35 and Appendix J.

CONTENT LOG

One row per post. Fill in reach, sends, saves, and watch time from the platform’s own analytics a few days after posting. Ask the door whether anyone mentioned it.

Fill in: Platform, Month.

Columns to fill in: Date, Format, Job it was meant to do, Reach, Sends, Saves, Avg. watch, Door mentions. Copy as many rows as you need.

Sends per reach = sends ÷ reach. Compare each post with your last eight of the same kind. Platforms define reach and watch time differently, so compare a number only with itself on the same platform.

MEASURE IT

The number: Sends per reach and saves per reach by post type, the retention curve for video, and guests at the door who name a post.

By hand: Log each post, then divide sends by reach and saves by reach. Compare room clips with graphics or any two formats. Count first-timers who name a post. Example: 81 ÷ 5,400 = 1.5% sends per reach. Compare with followers gained to see the gap.

Watch out: Platforms define reach and watch time differently and change them. One outlier isn’t a trend. Don’t chase sends into off-brand stunts.

Go deeper: Chapter 14 (content and social media) and Chapter 31 (running experiments in a real room).

ROLE BOX

• Owner: Ask the door which post, if any, brought each first-timer, and compare room clips with graphics every month.
• Booker: Track which acts’ clips travel, and share the numbers with the act. Use them in the proof file.
• Musician: Watch sends per reach and saves on live clips, and check where your viewers live before you route a tour.
• Bartender: Run the content log, read the retention curve, and compare followers with followers who visited.
• Dance Teacher: Log welcome, lesson, and spotlight posts. Count students who say a post brought them to their first lesson.

TALK ABOUT IT

• Which number on your platform dashboard do you look at first? What would you do differently if it dropped by half?
• How many of your followers could drive to your room tonight?
• Which kind of post gets applause and no guests? Which gets guests and little applause?