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

Raw view counts and follower deltas feel satisfying to screenshot, but they obscure the mechanics that determine whether a promotion budget was spent well. A campaign that delivers 80,000 views against a 100,000-view order looks fine until you notice that 60 percent of those views arrived in the first four hours and the rest trickled in over nine days. That pacing pattern tells you something about source quality that a single aggregate number never will.

Promotion analytics worth building a reporting practice around come down to three variables: delivery rate (did the order fulfill on schedule), survival rate (did the audience metric hold after delivery ended), and cost per thousand (what did each unit of reach actually cost once you normalize across packages). Every other number in a campaign report is either a derivative of these three or a vanity metric dressed up in a dashboard.

Delivery rate is the ratio of fulfilled units to ordered units over the contracted window. A 50,000-view TikTok package with a 72-hour delivery window should show a smooth ramp — not a spike on hour one followed by a flatline. When you pull delivery data from the promotion dashboard, look at the hourly or six-hour cadence, not just the end-of-campaign total. A 98 percent delivery rate that front-loaded 90 percent of volume in the first 12 hours is operationally different from a 98 percent rate spread evenly across the full window.

Front-loading is often a sign that the fulfillment layer is drawing from a single high-velocity source rather than a distributed one. That matters for content indexing and for how platforms weight early engagement signals. If your dashboard shows consistent front-loading across multiple orders, use the scaler to throttle the daily cap rather than accepting the default pace. Slowing delivery from 72 hours to five days on a 100,000-view order will, in most cases, produce a more stable downstream signal.

Document delivery rate per order in a simple log: date, package type, ordered volume, fulfilled volume, window length, and pacing shape (front-loaded, smooth, back-loaded). After ten orders you will have enough data to identify which package types consistently deliver clean pacing and which do not.

Survival rate measures what percentage of a delivered metric persists after a defined observation period — typically 7, 14, and 30 days post-delivery. A campaign that delivers 5,000 new followers and retains 4,200 after 30 days has an 84 percent survival rate. A campaign that delivers the same 5,000 and retains 1,100 has a 22 percent survival rate. These are not equivalent outcomes, and a client report that shows only the delivery number without the survival window is an incomplete picture.

To track survival rate without manual overhead, set a recurring audit task in your reporting workflow: pull the follower or metric count at delivery close, then again at day 7, day 14, and day 30. The delta between delivery close and day 30 is your net retention. On most well-sourced campaigns, you should see survival rates above 75 percent at the 30-day mark. Rates below 60 percent warrant a conversation about source configuration before reordering.

Survival rate also affects how you interpret CPM. A campaign with a $2.50 CPM and a 40 percent survival rate has an effective cost per thousand retained impressions that is more than double the headline number. Run this adjustment on every order before presenting results to a client or using the data to justify a budget increase.

A $180 order for 90,000 views and a $95 order for 35,000 views look completely different at face value. Normalized to CPM, the first is $2.00 and the second is $2.71. That gap matters when you are managing monthly retainers across multiple clients and trying to allocate budget to the highest-efficiency packages first. CPM normalization is the only way to make those comparisons honestly.

The standard formula is straightforward: (total spend / total units delivered) x 1,000. Apply it at the order level first, then roll it up to the campaign level, then to the client level. When you have 60 or 90 days of data in the promotion dashboard, you will start to see which service categories consistently hit a target CPM band and which drift above it. Use that data to set internal benchmarks — for example, 'no video view package should exceed $3.50 CPM on a standard 72-hour window' — and flag orders that miss the benchmark before reordering.

One common CPM mistake is calculating it on ordered volume rather than fulfilled volume. If a 100,000-view order fulfills at 91,000 views, your CPM is based on 91,000, not 100,000. Using the ordered number understates your true cost per unit. Always pull the fulfilled figure from the delivery report before running the calculation.

A reporting template that covers these three metrics does not need to be complex. A spreadsheet with one row per order and columns for: order date, platform, package type, ordered units, fulfilled units, delivery window, pacing shape, day-7 survival count, day-14 survival count, day-30 survival count, spend, and CPM (fulfilled) is sufficient for most agencies running up to 20 orders per month. Above that volume, pulling directly from the dashboard export and running the calculations in a connected sheet saves meaningful time.

For client-facing reports, present delivery rate and survival rate as a pair, not separately. A one-page summary showing 'ordered 50,000 views, delivered 48,750 over 72 hours (97.5% delivery rate), retained 41,400 at day 30 (84.9% survival rate), effective CPM $2.46' gives a client everything they need to evaluate the campaign without requiring them to interpret raw numbers. It also establishes a baseline that makes the next campaign's results directly comparable.

Review this data at the campaign level monthly and at the account level quarterly. Monthly reviews catch pacing or survival anomalies early enough to adjust active orders. Quarterly reviews reveal structural patterns — which platforms, which package sizes, and which delivery windows produce the most consistent results for a given content category — that should feed directly into how you configure future orders in the scaler.

The most operationally useful scenario in promotion analytics is when delivery rate, survival rate, and CPM point in different directions. A campaign can have a 99 percent delivery rate (fulfillment worked), a 35 percent survival rate (retention failed), and a headline CPM that looks competitive. If you only read delivery and CPM, you conclude the campaign was successful. If you include survival rate, you identify a source quality problem that will compound across future orders if not addressed.

Conversely, a campaign with an 88 percent delivery rate (slightly under-fulfilled), a 91 percent survival rate, and a CPM that came in above benchmark might still represent the better spend when adjusted for retention. The point is not that one metric outranks the others — it is that reviewing all three together gives you a complete diagnostic that neither confirms nor dismisses a campaign based on a single number.

When the three metrics disagree significantly, open a ticket or reconfigure the order parameters before reordering. The disagreement is data. Use it.