gtmskills.com
GitHubContribute
LinkedIn ads & ABM audit

Reference file

Metric definitions and the ad-count model

metric-definitions-and-the-ad-count-model.md

Metric definitions and the ad-count model

All metrics come from the ZenABM connector fields: impressions, clicks, landingPageClicks, cost (USD), engagements (and video fields). Compute per format and per ad; roll up for the account.

Core rates

CTR                    = clicks / impressions
CPC                    = cost / clicks
CPM                    = cost / impressions * 1000
effectiveCTRtoLP       = landingPageClicks / impressions
effectiveCPCtoLP       = cost / landingPageClicks          # n/a when landingPageClicks == 0
engagementRate         = engagements / impressions
  • Effective CTR to LP is the honest "did this ad actually send someone to the site" rate — it strips out likes, comments, profile clicks and other non-LP clicks. For a plain image/carousel/video ad it usually equals CTR; for TLAs it is much lower than CTR because most TLA "clicks" are profile/expand clicks, not link clicks.
  • Effective CPC to LP is what each real landing-page visit cost. This is the number that matters for pipeline, and it is where TLAs win big ($3.06 vs $13+ for image).
  • If landingPageClicks == 0, report effective CPC to LP as n/a (no link / no LP traffic), never as $0 or infinity.

RULE — grade image ads (and any link-driving format) on effective CPC to LP, not raw CPC. Raw CPC bills for every click including likes, reactions, profile and expand clicks that never reach the site. The honest cost of an image ad is cost / landingPageClicks (effective CPC to LP), compared to the $13.23 image benchmark. Lead with that number in the scorecard, the benchmark grade and the reallocation for Single Image, Carousel and Video. Raw CPC can be shown alongside as secondary, but the pass/fail call and any "this format is too expensive" verdict must be driven by effective CPC to LP. (TLAs are the one format where raw CPC and effective CPC to LP diverge hugely in the other direction — always report both for TLAs too.)

TLA effective metrics (called out in the report)

For Thought Leader Ads specifically, from the engagement/creative data:

TLA effectiveCTR   = landingPageClicks / impressions
TLA effectiveSpend = cost / landingPageClicks        # "effective cost per LP click"

Only TLAs that carry a link produce landing-page clicks. A TLA with no link has landingPageClicks == 0 and therefore no effective metric — say "no link, so no landing-page metric," don't score it as a zero.

Deltas (period over period)

For any metric m, with cur = this 30 days and prev = the previous 30 days:

deltaPct = (cur - prev) / prev * 100        # guard prev == 0 → report "new" instead of a %

Report direction words plainly: impressions/engagements/clicks up is good; spend down with output flat or up is increasing efficiency; spend up with output flat or down is decreasing efficiency. Always pair a volume delta with a cost delta so the verdict is honest (e.g. "clicks +18% while spend -6% — you got more for less, efficiency is improving").

Ad-count model (are you running the right number of ads?)

Reverse-engineered from ZenABM's LinkedIn Ads Count Calculator. For the audit, use the user's last-30-days spend as the budget and their actual CPC (not an assumption):

monthlyBudget      = last30DaysSpend
CPC                = actual account CPC (cost / clicks over the window)
clicksPerAdPerDay  = 3                                   # ZenABM planning default
dailyBudget        = monthlyBudget / 30
minDailyPerAd      = max(CPC * clicksPerAdPerDay, 20)    # hard $20/day floor
affordableAds      = max(floor(dailyBudget / minDailyPerAd), 1)

affordableAds is the number of click-driving ads this budget can fund at healthy delivery at once. Compare to the actual number of currently-serving ads (list_creatives, servingOnly). Text ads run on set-CPC and don't consume this budget — exclude them from the "actual" count for a fair comparison.

Verdict:

  • actualAds > affordableAds by a meaningful margin → too many — the budget is spread too thin, each ad is starved below the delivery floor and LinkedIn can't optimize any of them. Recommend consolidating to ~`affordableAds` and pausing the weakest.
  • actualAds < affordableAds by a meaningful margin → too few — there's headroom to launch more creatives / formats (lead with TLAs) and give the algorithm more to optimize.
  • within ~20% either way → about right.

Per-ad delivery health (per-ad daily spend = monthlyBudget / 30 / actualAds, or CPC * clicksPerAdPerDay for a single ad's planned spend):

  • < $25/day → Too Thin (LinkedIn can't optimize)
  • $25-49/day → Watch
  • ≥ $50/day → Healthy

Surface the tier next to the count so the "too many / too few" verdict is grounded in delivery, not just a ratio.

Allocation (reallocate the budget)

  1. Current split = spend share per format (spend_format / total_spend).
  2. Rank formats by value: primarily effective CPC to LP (lower is better) and deal influence (if CRM), with TLAs as the default value leader.
  3. Recommend moving budget from the worst effective-CPC-to-LP / lowest-influence formats into the best ones. Give an actual dollar move ("shift ~$X/mo from Video into TLAs"), not just "do more TLAs." Keep at least a small always-on presence in reach formats (image) unless the data screams otherwise.