What to measure when you build in public
When you build in public, measure the path from a post to a paying customer, not the reaction to the post. Track visitors per post, revenue per visitor by source, time to first purchase, the share of visitors who return, and which posts produced sales at all.
Source: x.com → /blog/setup-guide
3 visits over 12 days
Source: google.com → /pricing
Fig.A sale alert in Telegram: the source and the page, not just the amount. Sample data, not a real site.
Likes, replies and follower counts are easy to see and easy to enjoy. They are also weakly connected to revenue. A post can get 400 likes and send nobody to your site. A quiet update can send ten people, two of whom buy. You only find out if you measure the second kind of thing.
Why are followers a poor scoreboard?
Followers measure attention on a platform you do not own. Revenue measures whether that attention turned into a decision to pay. The two can move in opposite directions: a viral joke adds followers who never click, while a plain "here is what I shipped" post brings a handful of the right people.
None of this means engagement is useless. Replies tell you what resonates, and an audience helps over the long run. It just should not be the number you steer by when the goal is a business that pays you.
The five numbers worth a weekly look
- Visitors per post. Give every post its own tagged link (more on that below). Compare visitors per post, not impressions, because a click is the first real act of interest.
- Revenue per visitor, by source. Divide revenue by visitors for each channel. A source with few visitors and high revenue per visitor deserves more of your time than a big source that earns nothing. We go deeper on this in revenue per visitor, the metric that beats traffic.
- Time to first purchase. Do people buy in the same visit, or after three weeks of lurking? This tells you how long to wait before judging a post. The path to purchase view shows medians for days and visits before buying.
- Returning visitors. If the same people keep coming back, your updates are building familiarity. If everyone is new every week, you are renting attention one post at a time.
- Posts that produced a sale. Not likes, not clicks. A short list of the posts with a purchase behind them is the most useful list you will keep.
How do you tag posts so you can tell them apart?
Pick one naming scheme and never improvise. A scheme we like for builders:
| Field | Convention | Example |
|---|---|---|
| utm_source | the platform, lowercase | x, linkedin, reddit, newsletter |
| utm_medium | the format | post, thread, reply, comment |
| utm_campaign | date and short topic | 2026-11-pricing-change |
A full link looks like https://yoursite.com/?utm_source=x&utm_medium=post&utm_campaign=2026-11-pricing-change. Social apps often hide the referrer, and some in-app browsers cannot be told apart from a normal browser, so the tag is what keeps the post identifiable. Our notes on which tweet made a sale cover the X-specific details, and sources and attribution explains that UTM tags win over the referrer when both exist.
Example (made-up numbers for illustration)
Say you posted four updates in November and linked each one with its own campaign tag.
| Post | Visitors | Sales | Revenue | Revenue per visitor |
|---|---|---|---|---|
| Pricing change | 180 | 3 | $147 | $0.82 |
| Behind-the-scenes screenshot | 640 | 0 | $0 | $0.00 |
| Bug-fix changelog | 95 | 1 | $49 | $0.52 |
| Revenue milestone | 310 | 1 | $49 | $0.16 |
The screenshot post got the most traffic and earned nothing. The pricing post got a fraction of the visitors and earned the most. The decision this triggers: write more posts about the decision a buyer is making (price, tradeoffs, what changed), and treat pure behind-the-scenes posts as audience building, not sales.
Fig.Which posts and threads paid, not which ones got likes. Sample data, not a real site.
What decision should each number trigger?
- Visitors per post low across the board: the problem is distribution or the call to action, not the product. Change the link placement before changing the offer.
- Visitors fine, revenue per visitor near zero: look at the landing page the link points to. Send posts to the specific page that matches the post, not always the home page.
- Long time to purchase: stop judging a post after two days. Review it after the median buying window instead.
- Few returning visitors: add something worth returning for, such as a changelog or a recurring update.
Mark each post on your analytics chart as you publish it. Chart notes let you pin "Posted pricing thread" to the day, so a later jump in the numbers has its reason next to it.
When this does not apply
If you have no product to sell yet, revenue per visitor is zero by definition and the numbers above will mislead you. Before launch, measure something earlier: signups, waitlist joins, replies from your target customer. Also, with very small traffic, a single sale swings every ratio. Do not read a trend into three visitors.
At a glance
| Question | Number to watch | What it tells you |
|---|---|---|
| Is the post getting clicks? | Visitors per tagged link | Distribution and call to action |
| Does the traffic pay? | Revenue per visitor by source | Where to spend your time |
| How long until I can judge a post? | Median time to purchase | Review window |
| Is an audience forming? | Returning visitor share | Familiarity, not just reach |
| Which work earned money? | Posts with at least one sale | What to repeat |
If you want to see these numbers per post without building the spreadsheet yourself, the live demo shows the dashboard with sample data.
Two related reads: how to know which post made a sale and the weekly review routine.
FAQ
Should I stop sharing follower counts and likes?
No. They are fine as a morale check and a rough signal of resonance. Just do not use them to decide what to build or where to spend your week.
How many posts do I need before the numbers mean anything?
There is no magic number. A handful of tagged posts is enough to spot a clear gap, like one post earning ten times another. With a few visitors per post, treat results as hints and keep testing.
Do I need UTM tags if my analytics already shows the referrer?
For social platforms, yes. Referrers are often missing or generic, so the tag is the dependable way to tell two posts on the same platform apart.
What if my sales happen weeks after the post?
Then judge posts on a longer window and use an attribution model that suits it. Last non-direct credit is the default, and first touch credits the visit that found the buyer.