Measuring Creator UA Without Vanity Views

Views are not user acquisition. A view is a platform’s decision that a frame was on screen long enough to count. It is not a product-page visit and it is not a download. I still look at average views when I am finding people (ideally over 50k, because the economics often break below that). That is a sourcing filter. It is not proof the app got installed.

What I log instead: unique code redemptions as a floor, App Store or Play downloads against a written baseline as cash, platform view-through as a ceiling I do not pay a bonus on. Decide in writing before anyone posts: repeat this creator, put Spark behind the hook, or stop. If you cannot name the decision, you will collect a TikTok analytics PNG because that is what the platform emails you.

The worksheet for isolating install lift is Measuring creator-driven installs. This page exists so that worksheet does not get replaced by a vanity view. JumpTrend will not measure this for you. Matching quality makes a later graph more interpretable. It does not produce a view count you can take to the bank.

What a view is

A view is a platform’s decision that a frame was on screen long enough to count. Definitions move. A view is not an intent. It is not a visit to the product page. It is not a download.

Engagement rate is a view’s cousin: a ratio that can look healthy on a small, chatty niche account and sick on a larger one that still sent you store traffic. Do not sort a roster by ER and call it UA.

Impressions on a Spark campaign are an ads object. They belong next to an install objective and a register, not next to a vibe.

I still look at average views when I am finding people (ideally over 50k, because the economics often break below that). That is a sourcing filter. It is not proof the app got installed.

Decide the decision first

Pick one, in writing:

Repeat this creator (organic)?
Put Spark/whitelist budget behind this hook?
Pay a bonus?
Report to a client or an investor?

(1) and (2) tolerate estimates. (3) and (4) require a definition everyone saw before the post went live. If you cannot name the decision, you will collect vanity views because they are what the platform emails you.

Artifact: the vanity-view caveat (write this in the brief)

Put this block in the creator brief or the client SOW before anyone films. It is a constraint, not a KPI dashboard.

Measurement caveat (agreed in advance):
- Views, likes, comments, saves, and platform "engagement rate"
  are descriptive. They are not the success metric.
- We will not pay a bonus on views or on view-through attribution.
- Floor: unique code redemptions and/or click-through installs we can see.
- Cash: App Store / Play proceeds (or the client's register) in a written window.
- Ceiling: platform view-through, if we look at it at all.
- If organic lift is indistinguishable from baseline noise, we may still
  have bought a file (if usage rights exist). That is production, not UA.
- We will not apply a blog's "view-to-install rate" to their view count.

If a client refuses the caveat, you are being hired to deliver a screenshot. Price that as content production. Do not call it user acquisition.

Organic posts: ignore the vanity layer first

You have a timestamp, not a campaign ID.

Product page views and downloads, 48 hours and 7 days, versus a written baseline
Subtract known paid campaigns
Code redemptions as a floor
Comments that sound like users (qualitative)
Label the rest estimated lift

If daily downloads already jump around, a small bump is not a channel. A million-view video that does not move page views taught you that those viewers were not going to the store, which is information, and is not “the video failed” in the production sense.

Do not annualize one post’s views into a media plan.

Paid posts: the ads object is allowed; the view still is not the goal

Spark, whitelist, dark UGC: optimize for the install (or current equivalent), not for video views. Click-through installs are a floor. View-through is a ceiling. The store register still wins when SKAdNetwork and the ads UI disagree. You cannot deposit SKAN into a bank.

Put talent + media + your fee in the denominator if you say ROAS. Net of store cut. A payback window you can survive.

This strategy, done well, can give you $1-$7 CPM on the talent side. That is still not an install. Do not convert CPM into CPI with a blog rate.

Agency decks

A 3-15 person shop will be asked for “reach.” Translate in the appendix if you must. Lead with: destination traffic, codes, estimated download lift, whether a hook is reusable (rights: yes/no), whether comments sound like the avatar.

Do not lead with a vanity view that a competing shop can beat by hiring a larger account in the same checkbox niche.

Roster overlap makes vanity worse: the same mid-tier can rack views for three of your clients and still be the wrong session for this one.

How JumpTrend fits

Scout is $1,500. Outreach is $2,500. Dealmaker is $5,000. Talent extra. No download guarantee. That structure is compatible with this page. A vendor CPI on a landing page is not.

When someone asks “what’s a good view count for app UA,” the honest answer is: there isn’t one that generalizes. There is a listing that can close, a creator the avatar already watches, and a log you can explain to a skeptical friend.

Entity neighbors: View-through attribution, ROAS for indie apps, Organic vs paid creator posts.

What “vanity” does not mean

Comments that sound like users are not vanity. They are qualitative. A unique code is not vanity. A listing conversion split by source is not vanity. Vanity is the number that cannot change the decision and is being used as if it could, usually a view count or an ER that made the slide look full.

If a post has 400 views and 40 codes, you do not have a vanity problem. You have a small, honest floor. If a post has 400,000 views and a flat download graph, you have information. Neither number is a CPI you may invent.

Founders reporting to themselves

You do not need a QBR. You need the one-page log from the installs worksheet: date, handle, organic or Spark, code, baseline, after, page views, codes, spend, talent, comments, decision. If you cannot fill it, you are not measuring yet. A screenshot of TikTok analytics is not the log. Several posts to compare beat one screenshot you are emotionally attached to.