Skip to main content
Industries · Marketplaces

Find where supply, demand and trust are out of balance

DecisionBox reads listings, orders, fulfilment and reviews, and surfaces seller concentration, regional imbalances and the friction that quietly costs you repeat buyers.

CriticalHub-and-spoke imbalance

64% of orders ship cross-state. The Northeast has a 14:1 demand-to-seller ratio.

+76% freight, +7 days transit, −0.18 review points on every cross-state ship.

RecommendationRecruit 50+ sellers in the four highest-demand undersupplied states.

Built for VPs of Marketplace, Heads of Seller Growth, Heads of Demand, and COOs at two-sided platforms.

The cost of missed insights

The patterns hiding between supply and demand are the ones costing the take-rate.

Three benchmarks marketplace operators are quietly paying right now. Each one points to the same gap: the data already has the answer; nobody has the bandwidth to read both sides at once. Every quarter of waiting compounds into GMV you don’t recover.

Capacity gap
~67%

The share of global e-commerce GMV projected to flow through marketplaces by 2027. Concentration is rising; visibility into supply geography and seller cohorts hasn’t kept pace.

Source: a16z Marketplace 100, 2024
Concentration risk
Top 10%

Of sellers drive roughly 80% of GMV on most marketplaces. When the top of the long tail wobbles, the aggregate dashboard reads "healthy" until it doesn’t.

Source: a16z marketplace research, 2024
Liquidity pull
~70%

Of new marketplace buyers never make a second purchase. The acquisition spend was paid on both sides; the second-purchase trigger never fired. Habit-formation signals sit in the cohort data.

Source: Marketplace Pulse benchmarks, 2024
From real runs

From a finding to a plan

Three pairs DecisionBox surfaced in real marketplace discoveries. Each finding ships with a numbered recommendation attached — not a chatbot you have to drive. See how the autonomous discovery loop produces ranked, validated outputs.

CriticalSeller concentration

The top 5% of sellers generate 53% of GMV. Of the top 20 sellers, 8 are based in a single state. One seller alone holds 78% of one state’s GMV.

53% GMV from top 5% · 8 of 20 single-state · 78% one-seller dependency
P1Supply de-risking

De-risk the single-state dependency: proactively recruit 10+ sellers in the at-risk state. Cap any single seller at 40% of category GMV before next quarter.

−38pp concentration · 10+ sellers recruited · 90-day window
CriticalHub-and-spoke imbalance

64% of orders ship cross-state. The supply hub holds 64% of GMV but only 38% of demand. Cross-state shipments pay +76% freight, take +7 days, lose 0.18 review points.

64% cross-state · +76% freight · +7 days transit · −0.18 review pts
P1Geographic rebalance

Open a seller-recruitment campaign in the four highest-demand undersupplied states. Convert ~3K cross-state orders to same-state fulfillment with a 90-day onboarding bonus.

+12pp on-time · −R$10/order freight · +0.3 review pts on converted
MediumSatisfaction-masked retention

97% of customers placed exactly one order. Repeat rate fell 3.93% → 1.86% across four quarters. Satisfaction scores between repeat and one-time buyers are nearly identical: 4.13 vs 4.08.

97% one-time · 3.93% → 1.86% repeat rate · 4.13 vs 4.08 CSAT
P2Second-purchase trigger

Launch a day-21 second-purchase nudge for first-time buyers above the ICP-fit threshold. Pair with a category-adjacent recommendation from the buyer’s last order.

+2–4pp repeat rate · ICP-fit first-time cohort · 60-day window

Each pair above is from a real DecisionBox run. Source data is anonymized for the public site.

Use cases

For two-sided platforms, vertical marketplaces, and managed networks — all at once.

Same engine, same data, different analysis areas. Pick the surface that matters this quarter.

Where supply is short, where demand is unanswered.

The supply dashboard shows total sellers. The demand dashboard shows total searches. Neither shows the geography or category where the two don’t match. DecisionBox reads both sides against each other and flags the imbalance — the same shape of finding that drives growth-opportunity discovery.

Sample analysis areas
  • Demand-to-seller ratio by region and category
  • Search-to-listing match rate by query intent
  • Cross-state and cross-border fulfillment imbalance
  • Category-level supply gaps masked by category averages
  • New-seller onboarding throughput vs incoming demand
Why DecisionBox

Built for marketplace teams who need to read both sides at once.

Three things every ops review surfaces. Every DecisionBox deployment ships with all three.

01

Sees what nobody had time to look for

You’ll see findings the team didn’t have bandwidth to pull. DecisionBox decides what to investigate; you review. Not a chatbot you drive — a ranked backlog of supply gaps, take-rate drifts, and cohort breaks you didn’t have to write.

02

Every number you’d defend in an ops review

Every finding checked against your data by a verifier and a refuter, with its verdict shown.

03

Your data stays with you

Your credentials never leave your environment. DecisionBox reads what’s already there; nothing exfiltrates.

For your security and compliance review

What Marketplaces security teams ask first

How Enterprise handles it

Does our data move?

No. DecisionBox only reads it where it is, and can run inside your own environment, on-prem or air-gapped.

What can the AI see?

Only what you allow. Personal and sensitive data is hidden before the AI sees anything.

Which AI does it use?

The one your company approved, in your cloud account or on your own servers.

Can we audit it?

Every action is recorded: who, what, when and under which rule.

Videos

See DecisionBox at work

All videos
Product tourThe AI analyst that finds what nobody askedA two-minute tour: discovery, validation, Ask and reports on a real marketplace dataset.
DecisionBox for business · Episode 1How to read a finding, and how much to trust itThe numbers, the validation verdict and the evidence behind a finding, before you share it.
DecisionBox for business · Episode 4From finding to decision: read a recommendationThe expected impact, the steps, the findings behind it and its validation.

The next ops review question, already answered.

Let's run it on your data.