Building a category scorecard

Why a scorecard nobody signed is worthless

A category review without a signed category scorecard costs you the whole review, and on the $18.4M Sprouts refrigerated salsa set that is a full reset cycle spent arguing taste instead of grading a plan. I have watched forty minutes evaporate over whether a fermented-line expansion was "working," with no agreed number to settle it: I said the category's 6.2% dollar growth was the wrong kind, the buyer said the fresh SKUs I wanted to cut were fine, and neither of us could be proven right because we had never written down what winning meant. When there is no target on the table, the meeting becomes a contest of taste, and taste loses to whoever owns the shelf.

The scorecard is the fix. It is the short list of sales, margin, unit, and share targets that the buyer and the supplier agree on before anyone talks tactics. It converts the assessment from a description of what happened into a commitment about what should happen next. Step 3 tells you the refrigerated salsa set is up 6.2% dollars and 1.1% units. Step 4 turns that into "plus 4% dollars, hold unit share, plus 40 basis points blended margin" and gets both sides to sign it. Without that signature, step 8, the review, has nothing to grade, and every recommendation between here and there becomes a matter of opinion.

The KPI set that belongs on a scorecard

A scorecard is not a data dump. It is the handful of metrics that actually decide the category, each one there because it answers a question the buyer is graded on. The trap is loading it with everything the SPINS extract can produce. A scorecard with thirty rows is a scorecard nobody reads, and one nobody reads is one nobody signs.

Here is the working set for the Sprouts refrigerated salsa category. Every metric has a definition, a primary owner of the interest (shopper, retailer, or supplier), and a tag for whether it leads or lags. Leading indicators move first and predict; lagging indicators confirm after the fact.

The leading-versus-lagging split matters more than analysts treat it. Dollar and unit sales are lagging: by the time they move, the decision that moved them is a quarter old. Velocity, weeks of supply, and repeat rate lead, because a velocity dip or a repeat-rate slide shows up before it drags the dollar line down. A scorecard built only on lagging metrics tells you the reset failed after it is too late to steer. Put at least one leading metric on it so step 8 is not purely a post-mortem.

Two of these deserve a note because analysts skip them. GMROII (gross margin return on inventory investment) is the retailer's real profitability metric, not raw margin percent, because a 40% margin item that sits for eleven weeks earns the shelf less than a 28% margin item that turns every ten days. And weeks of supply is the leading twin of GMROII: it is the inventory signal that tells you a slow SKU is about to become a markdown before the margin line records the loss. For the definition of the broader discipline these sit inside, the category management glossary entry is the anchor.

Turning the assessment into targets

The assessment reads the category. The scorecard sets what to do about it. The bridge between them is where most scorecards go soft, because it is tempting to take the assessment number and just declare it the target. The category grew 6.2%, so target 6.2%. That is an extrapolation, and it ignores everything the units line is telling you.

Walk the salsa numbers. The category is up 6.2% dollars but only 1.1% units. Roughly five points of that dollar growth is price and mix, not real demand. So the honest question for the reset year is not "how do we keep growing 6%," it is "how much of that 6% is durable, and where is the real unit demand." That reframing produces very different targets.

The load-bearing line is "hold unit share." When a category's units are up only 1.1%, there is almost no real unit growth to capture. Chasing unit-share gains in a flat-unit category means one of two things: stealing share through price, which trains the shopper to wait for the deal and hollows out the margin target, or winning share the category did not actually create, which is a rounding-error fight over a static pie. Holding unit share while the category laps its price increases is the honest target. It says: we expect real demand to stay roughly flat, so we will defend our position and grow dollars through mix, not chase phantom unit volume.

This is the same discipline the secondary broth thread teaches. The shelf-stable broth set at Kroger might run up 4% dollars and down 1% units in 84.51 data, pure price. A broth scorecard that targets +4% dollars and +2% units is implicitly promising unit growth in a category that is shrinking in units. Hold unit share there too, and let the dollar target carry the plan.

The item-level scorecard, nested under the category

The category scorecard is the buyer's. But a supplier walks into that review with a second scorecard: its own item-level targets, and they have to nest cleanly under the category ones. If your item targets contradict the category targets, the buyer sees it immediately and your credibility is gone. If they support the category targets, you are handing her the plan.

Verde Fresca sits at $1.42M in the category, 7.7% dollar share, number four, with five SKUs. Two of those SKUs are the fermented line, growing 34% YoY; the three fresh SKUs are flat. Velocity is $42 per store per week against a category median of $31, and ACV is 78% at Sprouts. Here is what the item scorecard looks like, each line tied to the category target it serves.

Every one of those item targets earns its place by moving a category number. The fermented +25% funds the category's one real demand engine. Holding the flat fresh line respects the hold-unit-share target instead of starving it. Lifting velocity from $42 toward $50 and ACV from 78% to 85% grows Verde Fresca dollars through per-store productivity and distribution, the two honest levers when the category itself is not adding units. This is the scorecard I would sign my name to, because I could defend every number in the room without reaching for a metric that flatters my brand at the category's expense.

Note what is not on it: a unit-share grab. Verde Fresca could chase a full point of unit share by discounting the fresh SKUs, but that fights the category's hold-unit-share target and trains the shopper to wait for the price. The item scorecard has to respect the category scorecard, or the buyer treats the whole deck as a facings grab dressed up as category advice.

Leading indicators and the vanity-metric trap

The most common way a scorecard lies is dollar growth that is really price. Raw dollar growth is the number everyone reaches for because it is big and it is positive, and it is the number that hides the most. The salsa category up 6.2% dollars looks like a win until you set the 1.1% unit line next to it and see that five of those six points never touched real demand. A scorecard that reports dollar growth without unit growth beside it is a vanity scorecard, and it will pass a review it should have failed.

The defense is to pair every lagging metric with its leading or its decomposing twin.

Velocity is the leading indicator I trust most on a salsa scorecard, because it moves before dollars do and it is hard to fake. A brand can post a dollar-share gain purely from picking up thirty new Sprouts doors, which is real but one-time. Velocity holding at $42 while ACV climbs from 78% to 85% is the durable signal: the new doors are actually selling, not just stocking. Weeks of supply is the other early warning. A fresh SKU whose weeks-of-supply is creeping up is telling you it is about to become a markdown, one or two review cycles before the margin line records the loss.

Anti-patterns that sink a scorecard

Three failure modes show up in almost every weak scorecard I have reviewed.

Too many metrics. A scorecard with thirty rows is unusable, and thirty rows was never rigor. The buyer cannot hold thirty targets in her head during a reset, so she anchors on two or three anyway and the rest are decoration. Pick the five to eight that actually decide the category. For the Sprouts salsa set that is dollar growth, unit share, blended margin, fermented-segment growth, and velocity. Everything else is diagnostic backup, not a target.

Targets with no baseline. "Grow fermented" is not a target. "Grow fermented from its current 11% category share and +28% run rate to a +25% reset-year target" is. A target without the baseline it moves from cannot be graded in step 8, because you cannot say whether plus 25% was ambitious or a layup without knowing where it started. Every target line needs the number it starts from, or the review has no way to score it.

Margin-only or sales-only scorecards. A scorecard that tracks only blended margin pushes the buyer to over-cut assortment and starve the growth segments, because the fastest way to lift margin percent is to drop the low-margin volume drivers. A scorecard that tracks only dollar sales pushes toward promotional buydown that wins the quarter and loses the margin. The Sprouts salsa scorecard has to hold both: plus 4% dollars AND plus 40 bps margin, so neither can be gamed at the other's expense. The tension between the two is the point. It is what keeps the plan honest.

There is a fourth, quieter one: a scorecard that never links back to the category role. Refrigerated salsa is a routine category, so its scorecard should not carry traffic-driving or destination-category targets. Setting a "grow trips" target on a routine category is asking a metric to move that the category structurally cannot move. The scorecard has to fit the role the category was assigned in step 2, which is why the strategy chapter reads the role and the scorecard together.

Doing this in Scout

The grind in building a scorecard is not deciding the targets. It is assembling the baselines the targets attach to, and refreshing them every cycle so the numbers you sign are current. That means pulling the category dollars and units this year against last, your segment shares, your velocity against the category median, and your ACV, out of raw SPINS or Circana extracts, and rebuilding the same pivots every six months so the scorecard has a defensible baseline in every row.

Scout sits on that data and builds the baseline cuts a scorecard needs as a saved view. The dollar-and-unit growth line, the segment shares, the velocity-versus-median comparison, the ACV read, all refresh instead of getting rebuilt. So the two-minute task is confirming this cycle's baselines against last cycle's. The half-day task, assembling those baselines from scratch every review, is the one it removes. What Scout does not do is set the target. It will not tell you that "hold unit share" is the honest call over "plus 3% units," because that is a judgment about how much of the dollar growth is durable, and that judgment belongs to you and the buyer. Scout assembles the evidence the scorecard stands on. It does not sign it.

The short version

• A scorecard nobody signed is a review that cannot grade itself; the scorecard converts the assessment into agreed sales, margin, unit, and share targets before anyone talks tactics.
• Turn the read into honest targets, not extrapolations: for a category up 6.2% dollars and 1.1% units, that is +4% dollars, hold unit share, and +40 bps margin, with item targets nested cleanly underneath.
• Keep it to five to eight metrics, give every target a baseline, hold sales and margin together, and put at least one leading indicator (velocity, weeks of supply) on it so step 8 is not a post-mortem.

Related: How to run a category assessment · Category strategies, and when to use each