Key Data vs. Influencing Data: Why Most Agency Reports Drown You in the Wrong Numbers

Quick Answer

Key data is the small set of numbers that tells you whether your business is healthier this month than last (contribution margin per channel, blended CAC vs LTV, incremental revenue, cash conversion). Influencing data is everything else: impressions, CTR, CPM, ROAS by platform, frequency, audience overlap. Both have a place, but they are not the same thing, and most agency reports treat them like they are. That is why your dashboard has 47 metrics and you still cannot answer the question "is this making me money?"

The 47-Metric Dashboard Problem

Open the last report your agency sent you. Count the metrics.

If it is more than ten, someone is hiding something. Probably without meaning to.

Agencies pile on numbers because the platforms hand them over for free, and because more rows of green-up-arrows looks like more work. It is not. It is noise dressed as rigor.

I have seen this a hundred times. A founder gets a 12-tab Google Sheet every Monday. CTR is up, CPM is down, ROAS is "trending positively," and the engagement rate is 4.7%. Three months later they realize revenue is flat and ad spend is up 30%. The report said everything was working. The bank account said otherwise.

The report was full of influencing data. Almost none of it was key data.

Here is the test I run on any incoming dashboard during an audit: cover the page with my hand, ask the founder to recite from memory which three numbers tell them whether last month was a win or a loss. If they cannot do it, the dashboard has failed them. If the agency has trained them to recite "ROAS was 4.2 and CTR climbed twelve percent," the dashboard has trained them on the wrong things.

What Counts as Key Data

Key data answers one question: did the business get healthier?

For most eCommerce brands we work with, that comes down to four numbers.

Contribution margin per channel. Revenue minus COGS minus ad spend minus shipping and returns, broken out by acquisition source. This is the only metric that tells you whether each channel is paying for itself in real dollars. The formula is not complex, but pulling the inputs together is where most agencies stop. If COGS is 35%, returns are 8%, and shipping is 12%, a 3.0 ROAS channel is producing roughly a 14 cent margin on every revenue dollar. A 5.0 ROAS channel on the same product profile is producing about 30 cents. That gap is the entire reason to know contribution margin instead of platform ROAS.

Blended CAC vs new-customer LTV. Total marketing spend across all channels (including agency fees, creative production, and influencer spend) divided by net new customers acquired. Then compared against your honest 12-month LTV for a first-time buyer. If blended CAC is $42 and new-customer 12-month LTV is $96, you are at a healthy 2.3 LTV-to-CAC ratio. If LTV is $58 against the same CAC, the channel mix is barely breaking even after factoring in the cost of capital. We treat anything under 1.8x as a red flag, anything between 1.8 and 2.5 as workable but worth interrogating, and above 2.5 as room to push spend.

Incremental revenue. What would not have happened without the ad spend. This is the metric every platform has an incentive to inflate, and it requires real testing to estimate honestly. We cover the mechanics of incrementality testing in a separate piece, but the short version is that incremental revenue almost always lands somewhere between 40% and 70% of platform-attributed revenue. Until you have a number from a real holdout test, your contribution margin is an optimistic estimate.

Cash conversion. How fast the marketing dollar comes back as deposited revenue. A brand running 90-day customer payback is a fundamentally different cash position than a brand running 14-day payback, even at the same LTV-to-CAC ratio. Cash conversion is the metric that forces or limits how fast you can scale, often without showing up anywhere else on the dashboard. We track it as a rolling 60-day cohort: of every $100 spent in week one, how much landed in the bank account by week eight?

Notice what is missing. ROAS is not on the list. Neither is CTR, frequency, or any platform-specific metric. Those are inputs. The four above are outputs.

What Counts as Influencing Data

Influencing data is diagnostic. It tells you why the key numbers moved, not whether they moved in a direction you want.

CTR drops? That points to a creative problem. CPM spikes? That points to an audience or auction problem. Frequency runs high? That points to a saturation problem. Audience overlap rising? That tells you the same customer is getting served too many ads across too many ad sets, which inflates frequency without growing reach. All useful, but only when paired with the key data. By themselves they are weather reports for a flight you may not even be taking.

The right move is to lead the report with the four key numbers, then use the influencing data underneath to explain the why. Most agencies do the reverse, and then bury the key data so deep the founder never sees it.

Here is what that looks like in practice. Take a week where contribution margin on Meta dropped from 22% to 14%. The influencing data should walk you through the cause: CTR fell from 1.8% to 1.3%, CPM held flat, frequency climbed from 2.1 to 3.4, and the highest-spending ad set was a six-week-old creative. Now you have a story. The creative is fatigued, the same audience is getting hit too often, and the margin compression is the result. The fix is a creative refresh, not a budget cut. Without the influencing data, you would have made the wrong call.

The lesson is that influencing data is essential, but it works as a follow-on, not a lead. Lead with the key. Diagnose with the influencing. Make decisions on the combination.

Why Agencies Default to Influencing Data

Two reasons.

First, influencing data is easy. Every ad platform exports it in one click. Contribution margin requires you to connect ad spend to COGS to revenue to refunds, and that is messy work. It is the kind of work an agency does not want to do unless the client makes them. Tools like Triple Whale, Northbeam, and Polar Analytics have made this dramatically easier in the last two years, but you still have to wire up the data sources, agree on a COGS methodology, and decide how to handle returns. That setup usually takes a week or two of integration work that most agencies bill as a one-time project or skip entirely.

Second, influencing data is safer. If CTR is up and ROAS looks fine, the report writes itself. If contribution margin is down, the agency has to explain why and what they are going to do about it. That is a harder conversation. Most agencies would rather avoid it. The structural problem is that agency retainers reward continuity, not honesty, so the report that keeps the client comfortable is the report the agency ships.

We lead every Slaymaker report with the four numbers above. The platform metrics live underneath, used only to explain movement. If a client cannot tell from the first page whether they made or lost money, the report is broken.

How to Build the Key-Data Layer Even if Your Agency Won't

If your agency cannot or will not produce a key-data view, you can build a workable version yourself in about an afternoon using Shopify or your commerce platform plus a spreadsheet.

Start with a simple monthly worksheet. Pull your gross revenue by channel from Shopify's sales attribution report. Subtract platform-reported ad spend by channel. Apply your COGS percentage (use a single blended number to start, get fancier later). Subtract your average shipping cost per order multiplied by orders, and apply your return rate as a deduction. What you have left, for each channel, is a rough contribution margin. It will not be perfect. It will be directionally honest in a way platform ROAS is not.

Then add a second tab for blended CAC. Total monthly marketing spend in one cell, net new customers from the same month in another, divide. Compare against your 12-month new-customer LTV, which Shopify can estimate or you can pull from your customer database.

That is the minimum viable version. Once you have it, the conversation with your agency changes immediately. You stop arguing about whose ROAS number is right and start arguing about which channel is actually profitable. That second argument is the one worth having.

What to Ask Your Agency This Week

Send them this email:

Going forward, I want our weekly report to lead with contribution margin by channel, blended CAC vs new-customer LTV, and incremental revenue if we can model it. Platform-level metrics can stay in the appendix as supporting detail. I can provide COGS, shipping, and return rate inputs by end of day Friday. Can you have the new report format ready by the following Friday?

If they push back with words like "complex," "directional," or "platform-attributed," that is your answer.

If they ship the report on time, you have a partner who is willing to be measured on the right things. That is worth holding onto.

FAQ

Is ROAS key data or influencing data? Influencing. ROAS does not account for COGS, returns, or whether the revenue would have happened anyway. A 5x ROAS on branded search can look great and contribute almost nothing incremental.

How many metrics should a weekly report have? Four to six on the front page. The rest belongs in an appendix or a drilldown view.

What if my agency says they cannot pull contribution margin? They can. It takes work to wire up the data sources, but every agency we know that says it is impossible just means they have not done it before.

Does this apply to brands under $30K/month in ad spend? Yes. The numbers are smaller, but the question is identical.

What tools make the key-data layer easier to build? Triple Whale, Northbeam, and Polar Analytics all do most of the wiring for you. Native Shopify reports plus a spreadsheet works if you would rather not add a tool.