There are two separate lenses to read, analyze and gain insights from marketing metrics. One organizes them by where in the customer journey they sit, before the click, through engagement, at transaction, and beyond into loyalty. The other organizes them by what question they answer, what did it cost, how many did we get, how good were they. Neither lens is complete on its own. Read only by journey stage, you know where a number sits but not what kind of question it’s answering. Read only by Cost, Quantity, and Quality, you know what question a number answers but not where in the funnel it applies. The two frameworks aren’t alternatives. They’re the two axes of the same matrix, and most of the interesting diagnostic work in a marketing account happens at their intersection.
Every funnel stage produces its own Cost, Quantity, and Quality reading. Laid out together, the metrics from our two earlier articles form a single grid:
- Before the Click (Channel): Cost (CPM, CPC, CPI), Quantity (Impressions, Reach), Quality (CTR)
- Engagement (on-website/app): Cost (Cost per Session), Quantity (Sessions, Users), Quality (Clicks-to-Sessions, Bounce Rate)
- Transaction: Cost (CPL, CAC), Quantity (Leads, App Events, Purchases), Quality (Conversion Ratio, ROAS)
- Loyalty: Cost (Cost per Repeat Customer), Quantity (Repeat Purchases), Quality (Retention Rate, LTV)
Read across a row, you get a Cost/Quantity/Quality view of a single stage. Is this stage efficient? Does it have enough scale? Is what it’s producing any good? Read down a column, you get a journey-stage view of a single dimension. Is Cost climbing as a customer moves deeper into the funnel? Is Quality holding up from click to loyalty, or dropping off somewhere specific?
The value of the grid isn’t any single cell. It’s comparing cells against each other, across both axes at once. A metric that looks fine when read on only one axis can look broken the moment you check it against the other.
Take CTR, sitting in the Before-the-Click row, Quality column. Read alone, a strong CTR looks like a healthy signal. Read against the Quantity in the same row, Impressions and Reach, it’s still fine if reach is proportionate. But read down the Quality column, against Conversion Ratio at the Transaction stage, a strong CTR paired with a weak conversion ratio tells you the ad is earning clicks the rest of the funnel can’t convert. The Channel-stage number was never wrong. It just wasn’t answering the question everyone assumed it was answering.
The same applies to Cost. A cheap CPL, read alone, looks efficient. Read against Quality in the same row, Conversion Ratio, a cheap CPL paired with a weak conversion ratio means the low cost was bought by lowering the bar on who counts as a lead. Cost didn’t actually improve. It moved sideways into Quality, in the same stage, where it’s easy to miss unless you’re deliberately reading both.
This is the pattern a one-axis read almost always misses: a problem in one cell rarely stays contained to that cell. It shows up as an oddly good or oddly bad number one row down, or one column over, and only a matrix read catches the connection.
Visualizing the matrix: quantity, quality, and cost in one chart
Because each stage produces a Quantity number, a Quality number, and a Cost number, all three can be plotted on a single chart for that stage: Quantity on the x-axis, Quality on the y-axis, and Cost represented as the size of a bubble.
Plotted this way, a healthy campaign or channel sits in a predictable zone: reasonable quantity, reasonable quality, a bubble size proportionate to both. An anomaly stands out visually rather than requiring a line-by-line read of a spreadsheet: a channel with a large bubble (high cost) sitting far right (high quantity) but low on the chart (poor quality) is immediately visible as the outlier worth investigating, without needing to compare three separate columns of numbers to notice it.
This view is most useful run per funnel stage rather than across the whole account at once. A Channel-stage bubble chart surfaces ad-level anomalies, while a Transaction-stage version surfaces which channels are producing leads or purchases that aren’t worth what they cost, even if they looked fine earlier in the funnel.
Take four lead-generation campaigns running in parallel, all reporting into the Transaction row of the matrix. Three of the four sit in a similar range: moderate lead volume, moderate CPL, and a conversion ratio from lead to qualified opportunity in a healthy band. The fourth reports the lowest CPL of the group by a wide margin, and the highest lead volume. On a Cost-only or Quantity-only read, this campaign looks like the clear winner: cheapest and biggest.
Plotted on the Quantity/Quality/Cost chart for the Transaction stage, the same campaign tells a different story: it sits far to the right (high lead volume), but low on the Quality axis (poor conversion ratio), with a bubble size that reflects total spend rather than efficiency. It is generating a large number of cheap leads that convert to opportunities at a fraction of the rate the other three campaigns manage. Read as a single Cost metric, it wins. Read on the matrix, it’s the campaign quietly dragging down the account’s real output.
An e-commerce client had one paid social channel that consistently posted the best ROAS in their monthly reporting: a Transaction-stage, Quality-column number that made the channel look like the obvious place to increase budget.
Reading that channel across the full matrix told a fuller story. Its Quantity, at the Transaction stage, was small. The channel was producing a strong ROAS on a low volume of transactions, largely because it was reaching a narrow, already-warm audience segment repeatedly. At the Channel stage, Reach had been flat for months. The ROAS wasn’t wrong, but it described a channel that had already saturated its efficient audience and had little room left to scale without the number degrading.
A second channel, with a moderately weaker ROAS, showed the opposite pattern: strong Quantity at the Channel stage (growing reach, largely untapped audience) and Quality metrics at the Transaction stage that were solid, if not exceptional. Read on ROAS alone, it looked like the second choice. Read on the matrix, it was the channel with real headroom to grow. Budget moved from the “best number” channel to the “best matrix” channel, and overall revenue grew faster than continuing to fund the channel with the better single metric would have allowed.
As with both of the frameworks this article combines, none of this is meant to be read against an industry standard. What a healthy cell looks like, in any row, in any column, depends on your own business, channel mix, and history. The matrix doesn’t tell you what “good” is. It tells you where to look, and what to compare a number against before deciding whether it’s good at all.
The habit worth building from this article, on top of the two before it, is simple: before acting on any single metric, check its row and its column. A number that looks strong in isolation is only half-read until you’ve checked what it looks like against its neighbors on both axes.
Most accounts get optimized one metric at a time, which is exactly how a channel with the best headline number ends up being the wrong place to add budget. At 8 Spades, we read every metric on both axes, funnel stage and Cost, Quantity, Quality, before deciding where a rupee should go next. If your reporting has a “best number” that might not be the best channel, get in touch.