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

Most promotion dashboards surface a dozen or more metrics, but operators who run high-volume campaigns consistently return to three: delivery rate, survival rate, and cost per thousand (CPM). Everything else — shares, profile visits, secondary conversions — is context. These three are the signal.

Delivery rate tells you whether the volume you ordered actually reached the platform. Survival rate tells you whether that volume held over time. CPM tells you what you paid per unit of reach. Together, they answer the only question that matters before you reorder: did the campaign execute as specified, and at what cost per real impression.

The reason to start here rather than with engagement or follower counts is simple: downstream metrics are uninterpretable if you do not know whether delivery completed. A low engagement rate on a campaign with 60% delivery is a different problem than a low engagement rate on a campaign with 100% delivery. Conflating them produces bad decisions.

Delivery rate is the ratio of impressions or units delivered against units ordered, expressed as a percentage. If you ordered a 50,000-view package and the dashboard records 47,200 delivered by the campaign close, your delivery rate is 94.4%. That is a clean execution. Anything below 85% warrants a support query before reorder.

Where operators go wrong is treating delivery rate as a proxy for campaign quality. It is not. It measures whether the infrastructure fulfilled the order — nothing more. A campaign can deliver at 100% and still underperform against its audience-growth objective if the targeting parameters were wrong or the asset was weak. Keep these layers separate in your reporting.

Pacing within delivery also matters. A 50,000-view order delivered in 4 hours looks identical in aggregate to one delivered across 72 hours, but platform behavior differs substantially between the two. Check time-series delivery in your dashboard, not just the final total. Front-loaded delivery can cause platform-side suppression that shows up only when you look at post-campaign organic reach.

Survival rate measures how much of your delivered volume persists after a defined window — typically 30 days. If a campaign delivers 50,000 views and the 30-day count is 43,000, survival rate is 86%. Platforms periodically audit engagement and remove activity that fails their quality checks. Survival rate is your measure of how much of your investment survived that audit.

A survival rate below 80% is a flag worth escalating. It suggests either that the traffic source used for the campaign did not meet platform quality thresholds, or that the content itself triggered a review that pulled associated engagement. The former is a vendor problem; the latter is an asset problem. Your promotion analytics dashboard should let you distinguish them by cross-referencing delivery source tags against survival curves.

Practically, factor survival rate into your effective CPM calculation. If you paid $18 CPM on a campaign that retained only 75% of delivered views, your effective CPM against surviving impressions is $24. That changes the unit economics of the campaign meaningfully, especially if you are comparing across multiple service tiers or content formats.

Raw CPM — total spend divided by delivered impressions, multiplied by 1,000 — is the standard entry point for comparing promotion costs across campaigns or platforms. It is useful for benchmarking and for setting initial budgets. But raw CPM assumes all delivered impressions are equivalent, which they are not when survival rates vary.

Survival-adjusted CPM is calculated by dividing total spend by surviving impressions rather than delivered impressions. The formula is identical; only the denominator changes. Running this calculation across a campaign set often reveals that the lowest raw-CPM option is not the lowest effective-CPM option once attrition is priced in. This is the single most common misread in post-campaign reporting at agencies managing multiple concurrent buys.

Set a standard in your reporting workflow: every CPM figure you present to a client or stakeholder should be labeled as either raw or survival-adjusted. The label removes ambiguity and trains clients to ask better questions about quality rather than just volume.

A minimal promotion analytics stack for a single campaign needs four data points: ordered volume, delivered volume, surviving volume at 30 days, and total spend. From those four numbers you can derive delivery rate, survival rate, raw CPM, and survival-adjusted CPM. Everything else is supplementary until you have those four locked.

For agencies running more than five concurrent campaigns, pull these figures into a shared reporting template that updates weekly. Flag any campaign where delivery rate drops below 85% or survival rate drops below 80% for review before the next budget cycle. The scaler tool is most useful when you are making volume decisions based on clean historical data — scaling a campaign that has a 70% survival rate without addressing the underlying quality issue just amplifies the loss.

Clients who receive campaign reports structured around these three metrics consistently ask sharper follow-up questions and make faster reorder decisions. The reporting format itself signals operational competence. That matters for retention as much as the campaign results do.

A promotion dashboard is only as useful as the filters you apply to it. The default view — all campaigns, all time — is fine for an overview but useless for diagnosing a specific underperformer. When a campaign looks off, filter first by date range to isolate the delivery window, then by content format to rule out asset-level issues, then by service tier to check whether the problem is isolated to one traffic source.

Cross-filtering delivery rate against survival rate in a single view will surface the most common failure pattern: campaigns that delivered fully but retained poorly. This pattern almost always points to a traffic quality issue rather than a pacing issue, and it should be resolved at the service selection stage on the next order — not by adjusting the pacing of the same service.

Export the filtered view as a CSV before escalating any issue. Support teams and account managers resolve flagged campaigns faster when the data is already isolated and formatted. A screenshot of the dashboard is harder to act on than a timestamped CSV with delivery counts, survival counts, and spend per campaign line.