Sign in
Methodology — no black boxes

What we know, how we know it,
and what we don't know yet.

NABCA-grade account intelligence costs six figures a year. The POD Index rebuilds ~70% of its practical value from public data — state liquor commissions, open sales files, menus, and a demand model validated against real store-level sales. Here's the honest inventory.

—
Licensed accounts tracked
2,991,321
Store × product placements indexed (2.86M DISCUS-categorized)
—
Accounts with verified intel
16
Control states with data today

How we compare to NABCA

NABCA is the industry's gold standard, and priced for the giants. Rough math: about a third of the account coverage at a small fraction of the cost.

NABCA subscriptionThe POD Index
CostSix figures a year$299 a state, $1,499 for all 16. One-time, no subscription
Accounts coveredEffectively all control-state accounts82,000+ licensed accounts, with brand-level data on about a third
Off-premise store demandMeasured depletionsMeasured where states publish store sales; modeled elsewhere. Predictions tested 42% better than market-average guessing on real sales
On-premise (bars)Bar-level depletions via distributor reportingBrands matched from captured menus on 8% to 45% of venues, depending on the state. A targeting signal, not depletions
How you get itContract and onboardingPay by card, sign in, download everything or use the matching and territory tools

We're not a NABCA replacement for a national supplier. We're the version of it a craft brand can actually afford, and we show you the gap before you pay.

How the engine works

1 · Collect

Every licensed account from 17 state commissions. Store-level assortments and sales wherever states publish them.

2 · Normalize

2.86M product placements mapped to DISCUS categories and 6 price tiers — the industry-standard language.

3 · Model

Iowa publishes true per-store sales. We train on it: demographics + assortment → demand mix per category × tier.

4 · Validate

Tested on 1,817 Iowa stores the model never saw: predictions were 42% more accurate than assuming every store sells the state average — we capture about half of what makes each store's demand different.

Off-premise — retail liquor stores

Three intel depths: demand-scored we predict what sells there by category × tier · assortment we know every SKU on the shelf · directory the account exists on your map, intel pending.

Where states publish real store sales (Iowa fully; Oregon, Michigan, Pennsylvania, Montana, Alabama at coarser grain) our model is calibrated against them, not guessing. Utah's predictions are recalibrated to DABS' published statewide category totals.

On-premise — bars & restaurants

Bar-level depletions exist only inside paid distributor-reported feeds like NABCA's — there's no public source. Our signal is what venues actually print: menu-verified brand-level intel extracted from captured menus · directory licensed venue, menu not yet captured. Read each bar as nested subsets — verified menus sit inside captured menus, inside every licensed account (Ohio: 9,053 accounts, menus for 6,492, verified 3,123).

What we don't have

NABCA's distributor-reported bar depletions (that data exists, behind their subscription), per-store sales in states that don't publish them, and menus for roughly two-thirds of on-premise venues. New Hampshire is not available yet. Every purchase is a dated snapshot: the date is shown on each state before you buy, so you always know how fresh your copy is.

Pick your states →