May 8, 2026 · 8 min read · Marketplace, KPI, Monitoring Plan
KPI monitoring guide for marketplace companies
Marketplaces have a monitoring challenge that other business models don't: they have two distinct customer groups — buyers and sellers — whose health metrics can diverge. A marketplace can look healthy on the buyer side while quietly losing supply. Or show healthy supply metrics while buyer conversion is silently degrading.
The second challenge is that marketplace metrics have indirect relationships that compound over time. A seller supply drop today doesn't hurt GMV today — it hurts GMV in 2-4 weeks when buyers can't find what they're looking for. By the time GMV drops, the supply problem is already months old.
This guide covers what to monitor in a marketplace, at what frequency, and — critically — how to monitor supply and demand health separately before problems materialize in GMV.
The two-sided monitoring problem
Most marketplace companies monitor GMV and conversion rates — the demand-side outcome metrics. But GMV is a lagging indicator. By the time GMV drops, the underlying supply or demand issue has already been building for weeks.
The leading indicators that matter are on both sides:
Demand-side leading indicators:
- ·Search-to-listing click rate (are buyers finding relevant listings?)
- ·Time on listing page (are they engaging with what they find?)
- ·Message/contact rate (are they initiating transactions?)
- ·Offer acceptance rate (are negotiations succeeding?)
Supply-side leading indicators:
- ·New listing creation rate (is supply growing or shrinking?)
- ·Seller response time (are sellers engaged?)
- ·Active listing ratio (what percentage of sellers are listing actively?)
- ·Listing quality signals (are listings complete and accurate?)
A complete monitoring setup watches both sides continuously and alerts when either degrades — before the effects appear in GMV.
The marketplace monitoring plan
| Metric | Frequency | Compare Period | Alert Threshold | Why It Matters |
|---|---|---|---|---|
| Gross Merchandise Value (GMV) | Hourly | Same hour, same day last week | Drop >20% | Primary health signal — this is the outcome metric everything else predicts |
| Transaction volume (completed) | Hourly | Same hour, same day last week | Drop >20% | Separates price effects from volume effects in GMV |
| Buyer-side conversion rate | Hourly | Same hour, same day last week | Drop >15% | Demand health — detects UX issues, trust issues, or checkout problems |
| New listing creation rate | Daily | Same day last week + 4-week average | Drop >20% | Leading supply signal — drops here hit GMV in 2-4 weeks |
| Active seller rate (sellers with ≥1 active listing) | Daily | 7-day rolling average | Drop >10% | Seller engagement health — a silent churn signal on the supply side |
| Search-to-transaction rate | Daily | Same day last week | Drop >20% | Funnel efficiency — measures whether buyers find what they're looking for |
| Average listing price | Daily | Same day last week | Drop >15% or spike >25% | Detects pricing anomalies, promotional effects, or supply quality shifts |
| Fraud/dispute rate | Hourly | Rolling 24h average | Spike >2× | Fraud campaigns and trust issues — marketplaces are a target for both buyer and seller fraud |
| Take rate (revenue / GMV) | Daily | 7-day rolling average | Drop >5% relative | Revenue integrity — an unexpected drop can indicate fee logic bugs or promo leakage |
| Seller NPS / satisfaction signals | Weekly | Same week last month | Drop >10 points | Seller retention leading indicator — unhappy sellers list less and eventually leave |
| Time-to-first-transaction (new sellers) | Daily | 7-day rolling average | Spike >20% | New seller onboarding health — longer time to first transaction predicts early seller churn |
| Repeat buyer rate | Weekly | Same week last month | Drop >10% | Trust and satisfaction signal on the buyer side |
Supply and demand have different time patterns
One of the most common mistakes in marketplace monitoring is applying a single comparison period to all metrics. Supply and demand have fundamentally different intra-week patterns.
Buyer activity often peaks midweek (Tuesday-Thursday) for B2B marketplaces and on weekends for consumer marketplaces. Buyer conversion rates follow this pattern — comparing a Tuesday to a Sunday is meaningless.
Seller activity often peaks in evenings and weekends for individual/prosumer sellers (when they have time to list), and during business hours for commercial sellers. New listing creation has a very different weekly shape than transaction volume.
Always segment your marketplace monitoring by supply metrics and demand metrics, and always use same-day-of-week comparisons.
Why supply monitoring requires longer lead times
The biggest mistake in marketplace monitoring is waiting for GMV to drop before investigating supply health. Supply degradation is slow and invisible until it isn't.
Here's the typical progression:
- ·Week 1-2: New seller acquisition drops 20%. Existing sellers are still active. GMV looks fine.
- ·Week 3-4: Active seller count starts declining as existing sellers run out of inventory or disengage. Listing creation rate drops. GMV starts softening slightly.
- ·Week 5-6: Buyers searching in specific categories find fewer listings. Search-to-transaction rate drops. Buyer satisfaction falls.
- ·Week 7-8: GMV drops noticeably. Leadership asks what happened. The answer: something that started 6 weeks ago.
With supply monitoring in place, you catch the Week 1-2 signal and investigate. With only GMV monitoring, you're managing Week 7-8 consequences.
Fraud in marketplaces: buyer and seller vectors
Marketplace fraud comes from both sides and requires separate monitoring:
Buyer-side fraud:
- ·Chargeback fraud (received item, claimed non-delivery)
- ·Account takeover (stolen accounts used for transactions)
- ·Fake dispute inflation
Signal: chargeback rate, dispute rate, refund rate spikes.
Seller-side fraud:
- ·Fake listings (bait-and-switch, non-existent goods)
- ·New account supply manipulation (bulk fake listings to game ranking)
- ·Identity fraud (using stolen seller identity)
Signal: new seller listing rate spikes, listing quality score drops, buyer dispute rate against new sellers.
Transaction fraud:
- ·Money laundering through legitimate marketplace structure
- ·Coordinated wash trading to inflate transaction volume
Signal: unusual transaction patterns (same buyer/seller pairs, round-number transactions, off-hours activity).
Monitor fraud_rate and dispute_rate hourly with a 2× rolling average threshold. Fraud campaigns move fast — by the time a daily review catches them, the damage is done.
Key marketplace patterns
Seasonal supply and demand cycles. Consumer marketplaces often see demand peak in Q4 and supply peak as people list unwanted items in January. B2B marketplaces follow industry-specific cycles. A 4-week rolling baseline handles most in-year seasonality; year-over-year comparisons are needed for annual cycles.
Category-level supply/demand imbalance. Aggregate marketplace health can mask severe imbalance in specific categories. A marketplace might show healthy aggregate GMV while one high-value category has deteriorating supply and declining buyer conversion. Monitor supply and demand by category, not just aggregate.
New seller cohort health predicts future supply. Track the percentage of new sellers (first 30 days) who reach their first transaction within 30 days. This is the "seller activation rate" — the supply-side equivalent of trial conversion. A drop in this metric predicts seller churn 60-90 days out.
Liquidity thresholds are nonlinear. Marketplaces have minimum viable liquidity thresholds in each category. Below that threshold, buyers can't reliably find what they're looking for and stop searching. Above it, the market is self-reinforcing. Monitor listing depth (number of active listings per category) to stay well above liquidity thresholds.
What good marketplace alerts look like
A supply drop alert:
S2 — New listing creation below baseline New listing creation: 1,840 today — down 23% vs same day last Tuesday (2,390) and down 18% vs 4-week average (2,240). Segment: All categories · Seller type: Individual sellers [Acknowledge] [Open seller dashboard] [Mark as known issue]
A fraud spike alert:
S1 — Dispute rate spike Buyer dispute rate: 3.8% — 4.2× the 24h rolling average (0.9%). Period: last hour · Segment: New sellers (0–30 days) · Category: Electronics [Acknowledge] [Escalate] [Review new seller queue]
The "new sellers, electronics" segment in the fraud alert is what makes it actionable. That's a classic fake listing or bait-and-switch pattern. Without the segment, it's an investigation that takes hours.
Starting point for new marketplace monitoring setups
Four metrics to start — two demand, two supply:
- ·GMV — hourly, same hour last week. Your primary outcome metric.
- ·Buyer conversion rate — hourly, same hour last week. Demand-side funnel health.
- ·New listing creation — daily, same day last week. Leading supply signal.
- ·Fraud/dispute rate — hourly, rolling 24h average. Trust and fraud health.
Once those are running, add the full supply/demand dual monitoring plan.
Lighthouse connects to your data warehouse and monitors marketplace metrics continuously — both buyer and seller side — with Slack alerts when either moves. Start for free →