Measuring UGC: from views to installs

views are entertainment, completion rate and saves are signal, installs are truth. measure UGC in that order and never let the metrics trade places, because a channel judged on views will optimize itself into a very entertaining way to acquire nobody. the attribution will stay partly broken no matter what you do, so the job is a measurement system honest enough to decide what to film next.

The metric hierarchy

installs sit at the top because they are the only number the business feels. below them, saves and completion rate, because they measure whether the content held and mattered to the person watching. below those, shares and buying-question comments. views sit at the bottom, and the reason is structural: view count is mostly a readout of how big a test batch the algorithm granted, which is luck and timing as much as quality. a 40,000-view video with weak completion and no saves is a firework. an 8,000-view video that people finished, saved and asked “what app is this” under is a business signal wearing a small number.

the ZuAI figure people ask about, 10K to 2M users at a $0.02 blended CAC, was a measured number, not a vibe, and the measuring mattered as much as the content. the arithmetic behind reading a blended CAC honestly is in the ZuAI CAC arithmetic.

Accept what link-in-bio cannot tell you

short-video attribution is broken by design and pretending otherwise wastes weeks. most platforms give organic posts no reliable click path: one bio link, in-app browsers that strip parameters, and the store handoff losing the trail even when someone does click. a large share of your real installs will come from people who watched, put the phone down, and searched your app’s name in the store an hour later. no dashboard will ever hand those back to you, so the system below is built to work without them.

UTM discipline on the links you do control

where a link exists, make it carry information. a separate UTM-tagged link per platform at minimum, per campaign where the surface allows it, and never one naked homepage link doing duty everywhere. the discipline is boring: consistent source and medium names, lowercase, decided once and never improvised, so that three months of data adds up instead of fragmenting across “tiktok”, “TikTok” and “tt”. tagged links will still undercount. their job is not the full truth, it is the relative comparison between platforms and campaigns, which they deliver as long as the tagging never drifts.

Spike matching is the honest attribution model early

match install spikes to posting times in your app store console. it is crude, and at small scale it is also the most truthful tool you have: post, then watch the next 24 to 48 hours of installs against your baseline. a video that moved anything shows up as a bump your daily install chart cannot hide. keep a simple log of what posted when, and the overlay of that log on the install curve becomes your real attribution report. when two videos post the same day, you lose resolution, which is a reason to space your strongest concepts, not a reason to buy attribution software before you have the install volume to justify it.

The weekly review that decides what to film next

the whole measurement system exists to feed one 30-minute meeting with yourself, once a week:

1. pull the week’s videos with completion rate, saves and the install bumps from your posting log.
2. mark each video against your own account’s median, not against last week’s outlier and not against other accounts.
3. pick the 2 or 3 concepts that earned variants, and write next week’s hooks for them.
4. kill the concepts that have now failed across a full batch, and note why in one line, so dead ideas stay dead.
5. sanity-check the volume itself, because the count of shipped videos is a metric too, and the volume math says it is the one you control most directly.

if you want the downstream numbers computed instead of argued about, the CAC and payback calculators in the free tools do that division for you.

measurement is the steering wheel of the engine in the UGC content-led growth playbook. the breakdown call is free. Let’s talk Growth.