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

Operators who buy promotion regularly tend to obsess over the front end — package size, niche targeting, delivery window — and then do almost nothing with the data that comes back. That gap is where budget gets wasted. If you cannot answer 'how much of what I bought actually hit the audience, and at what cost per thousand?' within 24 hours of a campaign going live, you are flying without instruments.

Promotion analytics is not a post-mortem exercise. It is a real-time control layer. The three metrics that carry the most operational weight are delivery rate, survival rate, and cost per thousand (CPM). Each one tells you something different about whether your spend is working, and each one has a specific intervention point if the numbers are off.

Delivery rate is the percentage of your purchased volume that has been distributed within the contracted window. If you ordered a 50,000-view TikTok package over 72 hours and your dashboard shows 18,000 views at the 36-hour mark, your delivery rate is 36 percent — roughly half of where it should be at the midpoint. That is a flag, not a final verdict, but it warrants a check.

Pacing matters here. A front-loaded delivery (70 percent in the first 24 hours of a 72-hour window) can look alarming but is sometimes intentional for campaigns trying to hit an algorithmic threshold fast. A back-loaded delivery with nothing in the first 48 hours is almost always a supply problem. The promotion dashboard gives you hourly pacing curves; learn to read the slope, not just the endpoint.

The intervention threshold most operators use: if delivery is more than 20 percent behind pace at any midpoint, open a support ticket or adjust the targeting parameters before more of the budget clears. Waiting until the campaign ends to notice under-delivery means you have already paid for units you did not receive.

Survival rate is the percentage of distributed units — views, followers, plays, or whatever metric your campaign is purchasing — that remain on the target asset after a defined holding period, typically 30 days. A 10,000-follower campaign with a 94 percent survival rate at day 30 retained 9,400 followers. The same campaign with a 71 percent survival rate retained 7,100. The difference in effective CPM between those two outcomes is substantial.

Low survival rate is almost always a quality signal, not a platform signal. When platforms run enforcement sweeps, low-quality inventory disappears in clusters — you will see the rate drop sharply over 48 to 72 hours and then stabilize. Gradual, continuous decline over weeks usually indicates that the delivered audience segment has low engagement affinity with the content, which is a targeting problem rather than an inventory problem.

Track survival rate per campaign in your reporting stack, not as a blended average across all activity. A blended number obscures which service tiers or targeting configurations are performing. Segment it by content type, platform, and package size to find the configurations worth repeating.

Cost per thousand (CPM) normalizes spend across campaigns of different sizes and on different platforms, making it the only metric that lets you do apples-to-apples comparison. A 500,000-view YouTube campaign at $180 total spend is $0.36 CPM. A 20,000-view Instagram Reels campaign at $22 total spend is $1.10 CPM. Neither number is inherently good or bad — the right CPM depends on your niche, the audience quality you need, and what that audience does after exposure.

Where CPM becomes analytically powerful is when you combine it with survival rate. Adjusted CPM — calculated as spend divided by surviving units at day 30, multiplied by 1,000 — is the number that actually reflects what you paid for durable audience presence. A $0.40 raw CPM with 68 percent survival produces an adjusted CPM of $0.59. A $0.70 raw CPM with 97 percent survival produces an adjusted CPM of $0.72. The cheaper package was not cheaper.

Set a CPM ceiling before you scale. When you use the scaler to increase volume on a campaign that is performing, the per-unit cost sometimes shifts due to inventory availability at higher volumes. Check the effective CPM on the first scaled increment before committing the full volume increase.

A reasonable reporting cadence for an active campaign looks like this: delivery rate checked at every 25 percent of the contracted window, survival rate pulled at day 7 and day 30, CPM calculated at end of delivery and recalculated at day 30 using surviving units. That is five data pulls per campaign, which takes about 15 minutes total if your dashboard is configured correctly.

For agency operators running multiple client campaigns simultaneously, the key habit is tagging campaigns by client, platform, and content type at the point of purchase. That tagging structure is what allows you to filter dashboard views later and produce segmented reports rather than aggregate summaries. Aggregate summaries hide underperformers. Clients ask why their specific campaign underdelivered; 'the overall portfolio performed well' is not an answer.

Reporting proof for a client is not a screenshot of a single metric. It is a delivery timeline, a survival snapshot at a defined holding period, and a CPM figure that contextualizes the spend. Those three elements together give a client something they can evaluate against their own business metrics.

Low delivery rate at midpoint: check targeting constraints first (overly narrow parameters reduce available inventory), then check for platform-side delays on the service status, then escalate. Do not add volume on top of an under-delivering campaign — that compounds the problem.

Low survival rate at day 7: pause additional spend in that configuration, pull the segment data to identify whether the drop is clustered (quality issue) or gradual (targeting issue), and run a smaller test with a different service tier before reordering at scale.

High adjusted CPM relative to your benchmark: revisit the package selection. Some service tiers carry a quality premium that is worth paying for certain content types and audiences, and is not worth paying for others. Use the historical CPM data in your dashboard to identify which configurations consistently fall within your target range and bias future buys toward those.