Promotion Analytics: How to Read Delivery, Survival Rate, and CPM Like an Operator

The default export from most promotion dashboards hands you impressions, reach, and a bar chart. Those numbers are not useless, but they answer the wrong question. The question operators need answered is not 'how many people saw this' — it is 'did delivery run at the right pace, how much of the audience stayed, and what did each thousand exposures actually cost me.' Three metrics. Everything else is commentary.

Promotion analytics becomes a decision tool the moment you stop reading it as a performance summary and start reading it as a delivery audit. A 50k-view TikTok package that completes in 18 hours instead of 72 tells you something went wrong with pacing — even if the raw view count looks fine. The dashboard number and the delivery shape are different data points, and conflating them is the single most common reporting error agencies make when presenting results to clients.

Delivery rate is the ratio of fulfilled impressions or views to contracted impressions or views, measured at any point during the campaign window. A 100k-view package checked at the 48-hour mark of a 96-hour window should be sitting near 50% delivery — not 90%, not 20%. Deviation in either direction is a signal worth investigating before the campaign closes.

Front-loaded delivery is the more common failure mode. When 80% of contracted volume lands in the first quarter of the campaign window, the audience growth curve looks like a spike followed by silence. That shape is readable to platform algorithms and to human audiences alike, and neither responds well to it. The delivery tab in a promotion dashboard should show you an hourly or sub-daily breakdown, not just a cumulative total. If you cannot see the shape of delivery, you cannot manage it.

Back-loaded delivery is less common but more damaging to reporting cycles. If a client's campaign reporting window closes before contracted volume has fully delivered, you are presenting incomplete data as final results. Set internal checkpoints at 25%, 50%, and 75% of the campaign window and pull delivery snapshots at each. That cadence catches problems while there is still time to escalate.

Survival rate — sometimes called retention rate or completion rate depending on the format — is the percentage of the audience that reached a defined endpoint: a video's midpoint, the 30-second mark, the full-view threshold. For a video promotion campaign, a survival rate below 40% at the midpoint usually means either the creative or the placement is wrong. The promotion did its job; something downstream failed.

The metric becomes more useful when you segment it by delivery source or placement type. A 50k-view package split across two placement types might show 62% midpoint survival on one and 31% on the other. That split tells you exactly where to shift volume on the next order — it is not a creative note, it is a buying note. Promotion analytics at this resolution requires your dashboard to log survival data alongside delivery data, not in a separate report that you manually reconcile.

Survival rate also functions as a proxy for audience quality. High delivery, low survival consistently points to inventory that is technically delivering impressions but not delivering attention. When you see that pattern, pull the delivery timestamp distribution. Volume that clusters between 2am and 5am local time in a target market will almost always show suppressed survival rates regardless of creative quality.

Cost per thousand (CPM) is the most quoted metric in promotion buying and the most abused. The number is meaningless without a definition of what qualifies as the thousand. A CPM calculated against raw delivery volume is a different number than one calculated against survived, in-target impressions — and that difference can be three to five times larger on a poorly structured campaign.

Operators should calculate at minimum two CPM figures for every campaign: gross CPM (total spend divided by total delivered impressions) and effective CPM (total spend divided by impressions that met the survival threshold). The gap between those two numbers is the cost of waste. On a well-paced, well-targeted campaign, the gap should be narrow. When it is wide, you have identified exactly where to push back on a supplier or restructure a future order.

Use the scaler interface to model CPM scenarios before committing to volume. If the effective CPM on a 100k-view package at a 55% survival rate comes out worse than a 60k-view package with tighter placement controls, the smaller package is the better buy. That math takes 90 seconds to run and is the difference between margin-positive and margin-negative campaign structures.

A client-facing promotion analytics report should answer three questions in order: Did the campaign deliver as contracted? Did the audience engage at the expected depth? What did each unit of real attention cost? That structure maps directly onto delivery rate, survival rate, and effective CPM — and it is the structure that makes a report defensible in a QBR or a budget review.

Pull your dashboard export into a single tab. Row one: contracted volume versus delivered volume, with the delivery shape (hourly or daily breakdown). Row two: midpoint survival rate versus benchmark. Row three: gross CPM, effective CPM, and the variance between them. Below that, you can add channel-level or placement-level breakdowns if the campaign ran across multiple surfaces. Everything else — screenshots, platform logos, reach graphics — goes in an appendix for clients who want it. The operative data is five rows and takes two minutes to read.

The benchmark numbers matter as much as the actuals. A 48% midpoint survival rate reads differently depending on whether the category benchmark is 35% or 65%. Maintain a benchmark library internally — by format, by platform type, by content vertical — and update it after every campaign. Over six months of consistent tracking, that library becomes the most valuable prompt you have for setting client expectations before a campaign starts rather than defending results after it ends.

When delivery rate is on track but survival rate is materially below benchmark, the problem is almost never the supplier — it is the creative-to-placement match. A high-production brand video dropped into a feed placement optimized for short-form content will show exactly this pattern: full delivery, collapsed survival. The fix is a placement adjustment, not a supplier complaint.

When survival rate is strong but CPM is running above model, the campaign is working but it is expensive. Before deciding to cut volume, check whether the effective CPM is still within the client's cost-per-outcome tolerance. A high CPM with strong survival and downstream conversion can still be the right buy. The number to optimize is not CPM in isolation — it is the ratio of effective CPM to outcome value.

When delivery rate lags and the campaign is past the 50% time window, escalate immediately through the dashboard's delivery monitoring view. Delayed volume that arrives in the final 20% of a campaign window compresses into a spike pattern, which reintroduces all the pacing problems described above. Getting ahead of delivery lag is a time-sensitive action, not a post-campaign note.