Promotion Analytics: How to Read Delivery, Survival Rate, and CPM Before You Scale

Raw view counts and follower deltas are vanity. They tell you what happened at the surface but not whether the delivery was healthy, whether the audience retained, or whether you paid a fair rate for real reach. Operators who rely on top-line numbers end up scaling broken campaigns because nothing in the summary flagged the problem.

Promotion analytics done correctly means tracking three interdependent signals: delivery rate, survival rate, and cost per thousand (CPM). Each one diagnoses a different failure mode. Delivery tells you if the traffic arrived on schedule. Survival tells you if the audience held after delivery ended. CPM tells you if the unit economics are defensible at the next volume tier. Miss any one of them and your reporting is incomplete.

Delivery rate measures what percentage of a purchased volume actually arrived within the contracted window. A 50,000-view TikTok package ordered over 72 hours should show continuous, graduated delivery across that period — not a spike on hour one followed by nothing. Uneven delivery is often the first sign of low-quality sourcing, oversold inventory, or pacing misconfigurations on the supply side.

On the promotion dashboard, delivery is most usefully viewed as an hourly or six-hour rolling curve rather than a cumulative total. Cumulative totals make a lopsided delivery look acceptable because the end number still hits the target. The curve reveals whether the traffic pattern matches what an organic audience would actually see — and whether the platform's own algorithm would treat it accordingly.

A healthy delivery curve for a 72-hour campaign looks like a gradual ramp over the first 12 hours, a sustained plateau through the middle window, and a natural taper at close. Any campaign where more than 60% of volume arrives in the first 24 hours warrants a manual review before reordering at higher volume.

Survival rate is the ratio of retained metrics — followers, subscribers, channel views, engagement counts — measured 7 to 14 days post-delivery versus the peak delivered number. A campaign that delivers 10,000 YouTube subscribers but retains only 6,200 at day 14 has a 62% survival rate. That number is your real acquisition figure, not the 10,000.

Low survival rates usually point to one of two problems: audience quality at the sourcing level, or a mismatch between the content and the audience profile delivered. Both are diagnosable. If survival is low across multiple campaigns on the same content, the sourcing needs to change. If survival varies sharply between campaigns on similar content, the targeting parameters are the variable to adjust.

Build survival rate into every post-campaign report as a standard line item. Reporting delivery volume without survival is like reporting gross revenue without churn — the number is technically accurate and operationally useless.

Cost per thousand impressions or acquisitions means nothing without a channel-specific baseline. A CPM of $4.20 for YouTube views is either excellent or poor depending on the niche, the geography of the audience, and whether the traffic is being delivered to a monetized channel where RPM matters downstream. Generic industry CPM tables are a starting point at best.

The right way to build a CPM benchmark is to run three to five campaigns at moderate volume — say, 25,000 to 75,000 units — across the same content vertical, log the delivery rate and survival rate for each, then calculate an effective CPM that accounts for survival. If you paid $3.50 CPM delivered but only 70% survived, your real cost per thousand retained units is $5.00. That is the number to carry into procurement conversations.

When volume scales — moving from a 50k package to a 250k package via the scaler — CPM should decrease modestly as sourcing efficiency improves. If CPM stays flat or rises at higher volume, that is a signal that the supply pool for that channel type is thin and the inventory is being sourced from more expensive tiers to meet demand.

A usable promotion analytics report has four sections: a delivery summary with the hourly curve exported or screenshotted from the dashboard, a survival snapshot taken at day 7 and day 14, a CPM table showing gross CPM versus effective CPM post-survival, and a one-paragraph diagnosis noting what to change on the next order.

The diagnosis paragraph is the part most reporting templates skip. It should answer three questions: Did delivery pace correctly? What was survival, and what likely caused any drop? Is the effective CPM within range for this channel at this volume? Answering those three in plain language turns a data export into an operational memo that informs the next budget decision.

For agency clients, present the effective CPM — not the gross CPM — as the headline cost figure. Clients who see a low gross CPM and then watch follower counts drift down over two weeks lose trust in the channel and in you. Leading with the number that accounts for retention is more conservative and more defensible.

If delivery rate is below 85% of contracted volume within the window, flag it before the campaign closes, not after. Most fulfillment issues are correctable mid-campaign if caught early — a pacing adjustment or a sourcing swap can bring the curve back in line. Post-campaign credits are a worse outcome for everyone.

If survival rate drops below 65% at day 7, do not reorder the same package. Run a smaller test at half volume with different targeting parameters — audience geography, age bracket, or interest category — before committing more budget. Use the scaler to cap the reorder volume until the revised configuration proves out.

If effective CPM is more than 30% above your established baseline for that channel type, trace the cause before scaling. It is usually one of three things: the content changed in a way that reduced relevance match, the sourcing pool shifted, or the campaign window was too compressed and the delivery rate suffered as a result. Each has a different fix.