Marketplace Analytics: KPIs Every Seller Should Track
A marketplace analytics framework covering native metric definitions, discovery, conversion, inventory, fulfilment, returns, advertising, contribution, cohorts and decision dashboards.
Marketplace analytics should explain what happened, why it matters, and what decision follows. A dashboard filled with views, likes, orders, and revenue can still mislead if platforms define those metrics differently or if sales lose money after fees, shipping, returns, and labour.
The solution has two layers:
1. Preserve each marketplace's native definitions for operational and channel analysis.
2. Build a normalized business layer for comparable orders, units, contribution, inventory, and customer outcomes.
This guide defines the key performance indicators, formulas, data model, reconciliation process, dashboard views, review cadence, and diagnostic questions a multichannel seller can use.
Key Takeaways
- Write a metric dictionary before combining platform reports.
- Do not treat views, visits, impressions, clicks, orders, units, revenue, and payouts as interchangeable.
- Preserve marketplace-native attribution and date logic.
- Normalize financial and inventory metrics in a separate business model.
- Measure the complete funnel: eligible inventory, impressions, visits/clicks, orders, fulfilment, returns, and contribution.
- Segment by channel, product, SKU, category, source, cohort, condition, and promotion.
- Compare paid and organic outcomes without assuming platform-attributed revenue is incremental.
- Reconcile orders, refunds, fees, shipping, and payouts.
- Use median and percentile operational measures, not only averages.
- Track inventory age and capital efficiency.
- Attach every KPI to an owner, threshold, and action.
- Protect buyer data and avoid collecting more personal information than needed.
Start With a Metric Dictionary
For every metric record:
| Field | Definition |
|---|---|
| Name | Exact label |
| Business question | What decision does it support? |
| Formula | Numerator and denominator |
| Source | Platform/report/table |
| Grain | Shop, listing, order, line, SKU, day |
| Date | Event date used |
| Time zone | Reporting boundary |
| Currency | Original and normalized |
| Attribution | Organic, ad, last click, platform-defined |
| Exclusions | Tax, shipping, cancellations, bots |
| Refresh | Real-time, daily, delayed |
| Owner | Who reviews and acts? |
| Threshold | Expected/alert range |
| Limitation | Known uncertainty |
Without a dictionary, two people can use “conversion rate” to mean different calculations.
Preserve Native Definitions
Etsy's current Shop Stats glossary defines:
- Visits as people who looked at the shop or listings
- Listing views as total listing views
- Orders as orders, not items
- Conversion rate as orders divided by visits
- Revenue under Etsy's specified calculation
Etsy also explains that one shopper viewing six listings can produce one visit and six listing views.
Do not replace Etsy's native conversion number with a different formula and label it “Etsy conversion rate.” Name your normalized version separately.
eBay Seller Hub provides sales, selling costs, traffic, traffic sources, seller level, payouts, and reports. Its Seller Hub documentation describes these sections, but exact fields may vary by account and marketplace.
Build the Measurement Funnel
Use six stages:
1. Sellable inventory
2. Discovery
3. Engagement
4. Conversion
5. Fulfilment and service
6. Economics and retention
A decline at one stage does not justify changing another.
Inventory KPIs
Active sellable SKUs
Count unique internal SKUs available for purchase.
Do not count the same SKU three times because it is cross-listed on three channels.
Active channel offers
Count marketplace listings/offers. Useful for channel coverage, not physical inventory.
Inventory value at cost
`text
Inventory value =
sum(on-hand units × allocated unit cost)
`
Choose cost basis with accounting guidance.
Inventory age
`text
Inventory age =
review date - available-for-sale date
`
Use available-for-sale date rather than acquisition when preparation varies, but track both.
Sell-through
Define the denominator explicitly.
`text
Cohort sell-through =
units sold from cohort
/ units initially available in cohort
× 100
`
For marketplace research tools, retain the platform's definition separately.
Stockout rate
`text
Stockout rate =
sellable SKU-days unavailable
/ planned sellable SKU-days
× 100
`
For one-of-a-kind inventory, stockout may mean a successful sale; interpret by assortment strategy.
Inventory accuracy
`text
Inventory accuracy =
correct counted SKU-location records
/ counted SKU-location records
× 100
`
Also track channel quantity mismatches and oversell incidents.
Discovery KPIs
Impressions
Times a listing or ad was shown under the platform's definition.
Use for visibility, not audience size. One person can create many impressions.
Search impressions
Separate marketplace search from browse, recommendations, ads, social, and external traffic when data allows.
Search-query coverage
Count qualified queries producing visibility or visits for a defined listing group.
Do not reward irrelevant query expansion.
Impression share
Use only where a platform provides a valid denominator. Do not invent total market impressions.
Eligible listing rate
`text
Eligible listing rate =
listings without known policy/data visibility blocks
/ active listings
× 100
`
Track required attributes, search visibility notices, and suppressed listings.
Engagement KPIs
Click-through rate
`text
CTR =
listing clicks or visits from impression
/ eligible impressions
× 100
`
Use the platform's numerator and denominator when reporting native CTR. “View” may mean a listing page view, not a search click.
Listing views per visit
Etsy exposes this as a native ratio. It can indicate shop exploration, but more views may also mean buyers cannot find a suitable product.
Saves, likes, favourites and watchers
Useful as intent signals, not revenue. Segment by eventual sale and inventory age.
Question rate
`text
Pre-sale question rate =
qualified product questions
/ listing visits or orders
`
Choose denominator based on use. Categorize questions to improve missing information.
Conversion KPIs
Marketplace-native conversion
Use the platform's displayed definition.
Normalized order conversion
When comparable sessions/visits exist:
`text
Normalized order conversion =
orders
/ qualified visits
× 100
`
Do not compare channels if visit definitions materially differ.
Unit conversion
`text
Units per visit =
units sold
/ visits
`
Useful when bundles or multi-item orders are common.
Orders
Deduplicate by marketplace order ID. Track status: pending, paid, cancelled, fulfilled, refunded.
Units per order
`text
Units per order =
net units sold
/ completed orders
`
Average order value
Define inclusions.
`text
AOV =
item revenue for completed orders
/ completed orders
`
State whether shipping, tax, discounts, refunds, and fees are included.
Offer acceptance
`text
Offer acceptance =
accepted offers
/ eligible offers received
× 100
`
Also track accepted-price discount and contribution.
Revenue Is Not Payout or Profit
Separate:
- Gross merchandise value
- Item revenue
- Shipping collected
- Discounts
- Tax collected
- Refunds
- Net sales
- Marketplace fees
- Payment processing
- Advertising
- Shipping labels
- Other adjustments
- Payout
- Product cost
- Labour and overhead
- Contribution
A payout can include timing adjustments from several orders and refunds. It should reconcile with transactions but is not an order-level profit metric.
Contribution KPIs
Contribution per order
`text
Contribution =
item revenue
+ shipping collected
- discounts/refunds
- product cost
- marketplace/payment fees
- advertising attributed cost
- packaging
- outbound shipping
- return/loss allowance
- variable preparation and fulfilment labour
`
Contribution margin
`text
Contribution margin =
contribution
/ applicable net revenue
× 100
`
Define revenue consistently and handle negative/refunded orders.
Contribution per listing hour
`text
Contribution per labour hour =
contribution
/ sourcing, preparation, listing and fulfilment hours
`
This reveals products that sell but consume unsustainable manual effort.
Profit after fixed cost
Contribution is not final profit. Subtract fixed payroll, rent, insurance, subscriptions, professional services, and other operating costs according to the accounting model.
Advertising KPIs
Ad impressions and clicks
Preserve native definitions by campaign type.
Ad click-through rate
`text
Ad CTR =
ad clicks / ad impressions × 100
`
Attributed orders/revenue
Use the platform's attribution window and rules. Attributed does not necessarily mean incremental.
Return on ad spend
`text
ROAS =
attributed revenue / ad spend
`
ROAS ignores product cost, fees, returns, and labour.
Advertising contribution
`text
Advertising contribution =
attributed order contribution before ads
- ad cost
`
Incrementality
Use holdouts, comparable groups, or controlled periods when possible. A buyer who would have purchased organically may still be attributed to advertising.
For boosted Depop listings, account for the current view-based attribution window and fee rules explained in the Depop Pricing Strategy guide.
Fulfilment KPIs
Handling time
Track order-ready to carrier acceptance, not merely label creation.
Use median and 90th percentile.
On-time dispatch
`text
On-time dispatch =
orders accepted by carrier within promise
/ eligible shipped orders
× 100
`
Tracking coverage
`text
Valid tracking coverage =
eligible orders with valid tracking
/ eligible shipped orders
× 100
`
Delivery issue rate
`text
Delivery issue rate =
lost, delayed or damaged delivery cases
/ shipped orders
× 100
`
Segment by carrier, service, package type, destination, and product.
Fulfilment cost per order
Include label, packaging, insurance, handling labour, and adjustments.
Returns and Service KPIs
Return request rate
`text
Return request rate =
orders with return request
/ eligible delivered orders
× 100
`
Refund rate
Measure orders and value separately.
`text
Refunded order rate =
orders with full or partial refund
/ completed orders
× 100
`
Case/dispute rate
Use platform definitions for seller standards, plus an internal normalized count.
Initial response time
Track median and 90th percentile. Platform customer-service standards may use a particular message definition.
Root-cause distribution
Categorize:
- Size/fit
- Condition
- Colour/material
- Compatibility
- Missing item
- Not as described
- Delivery
- Damage
- Buyer preference
- Fraud/abuse
- Seller cancellation
- Inventory mismatch
The root cause determines the action.
Seller Quality KPIs
Track native account standards separately:
- Review rating
- Message response
- Cases
- Defects
- Late shipment
- Tracking
- Cancellation
- Policy warnings
- Listing suppressions
Do not average different platform quality scores into one meaningless percentage.
Create a red/amber/green control view:
- Red: action or account risk
- Amber: deteriorating trend
- Green: within operating range
Product and Listing KPIs
At SKU/listing level:
- Impressions
- Visits/views
- CTR
- Saves/watchers
- Orders
- Units
- Conversion
- Accepted price
- Discount
- Contribution
- Days to sale
- Questions
- Returns
- Ad cost
- Inventory age
Classify listings:
| Segment | Interpretation |
|---|---|
| High visibility, high conversion | Protect stock and operations |
| High visibility, low conversion | Offer/listing problem |
| Low visibility, high conversion | Discovery opportunity |
| Low visibility, low conversion | Relevance, demand, or retirement review |
Avoid changing all listings in a segment without manual inspection.
Cohort Analysis
Group inventory by:
- Acquisition month
- Listing month
- Source
- Product type
- Brand
- Condition
- Price band
- Channel
- Seller process version
- Promotion
- Season
Track how a cohort sells over 30, 60, 90, and 180 days.
Cohorts reduce survivorship bias. Looking only at current active inventory hides products already sold or retired.
Cross-Channel Normalization
Create normalized fact tables:
Order
- Internal order key
- Marketplace
- External order ID
- Order date/time
- Currency
- Status
- Item revenue
- Shipping
- Tax
- Discount
- Refund
- fees
- payout reference
Order line
- Order key
- Internal SKU
- External listing/variation ID
- Quantity
- Unit revenue
- Allocated discount/refund/fees
- Product cost
- contribution
Listing snapshot
- Date
- Channel
- Listing ID
- SKU
- status
- price
- quantity
- impressions
- visits/views
- favourites
- ad state
Inventory event
- SKU
- location
- event type
- quantity
- source
- timestamp
- reference
Keep raw exports immutable and transform into normalized tables.
Currency and Time Zones
Store original currency and transaction amount. Normalize using a documented exchange-rate source and date for management reporting.
Do not compare revenue across currencies without conversion.
Use:
- UTC timestamps internally
- Marketplace/local timestamp retained
- Business reporting time zone
- Clear day boundary
Orders around midnight can shift periods when reports use different time zones.
Reconciliation
Daily
- New orders
- Cancelled/refunded orders
- Unmapped SKUs
- Failed inventory updates
- Shipping labels
- High-risk account alerts
Weekly
- Orders vs transaction report
- Refunds and fees
- Advertising charges
- Listing suppressions
- Inventory mismatch
- Old unfulfilled orders
Monthly
`text
Opening receivable
+ sales and shipping
- refunds
- fees and labels
+/- adjustments
- payouts
= closing receivable
`
Investigate differences; do not force a balancing adjustment without explanation.
Dashboard Design
Executive view
- Net sales
- Contribution
- Contribution margin
- Orders/units
- Inventory at cost
- Sell-through
- Return/cancellation
- Account-risk alerts
Growth view
- Qualified impressions
- CTR/visit
- Conversion
- AOV
- Organic vs paid
- Query/category performance
- Listing segments
Operations view
- Unfulfilled orders
- Dispatch performance
- Inventory mismatches
- Returns
- Cases
- Integration failures
- Old inventory
Sourcing view
- Contribution by source/category
- Days to sale
- Cohort sell-through
- Return risk
- Labour
- Maximum-buy feedback
Every chart should show definition, period, filter, currency, source and refresh time.
Review Cadence
Daily
Exceptions, orders, inventory, fulfilment and account risk.
Weekly
Funnel changes, priority listings, ads, returns and aged inventory.
Monthly
Financial reconciliation, contribution, cohorts, channel mix and sourcing rules.
Quarterly
Strategy, platform dependency, category economics, tool cost, data quality and targets.
Do not wait for a monthly meeting to address a listing suspension or oversell risk.
Statistical Caution
- Small samples fluctuate.
- Seasonality invalidates naive prior-period comparison.
- Average can hide extremes.
- Correlation is not causation.
- Platform attribution can overlap.
- Delayed refunds revise prior periods.
- Bot filtering can change traffic retrospectively.
- Changing several variables destroys attribution.
- Survivorship bias hides failed inventory.
- Revenue growth can reduce profit.
State uncertainty and minimum sample expectations.
KPI Ownership and Action Rules
A metric without an action is decoration.
Example:
| KPI | Threshold | Owner | Action |
|---|---|---|---|
| Inventory mismatch | Any sold SKU active | Operations | Disable listing and investigate |
| Contribution/order | Below floor | Finance/pricing | Audit price and cost |
| 90th percentile dispatch | Above promise | Fulfilment | Reduce handling promise/volume |
| Return reason: size | Above baseline | Catalog | Review measurements |
| Paid contribution | Negative after test | Marketing | Pause campaign |
| Unmapped order line | Any | Data/operations | Map before fulfilment |
Keep thresholds specific to category and maturity.
Data Quality Controls
Validate:
- Unique order IDs
- No duplicate lines
- SKU mapping
- Quantity signs
- Currency
- Time zone
- Refund allocation
- Fee completeness
- Missing cost
- Negative inventory
- Payout reconciliation
- Snapshot gaps
- Report refresh
- Platform schema changes
Display data-quality status beside KPIs.
Privacy and Security
Marketplace exports can contain personal data.
Apply:
- Least access
- Protected storage
- No unnecessary buyer data in analytics
- Retention and deletion
- Secure sharing
- Redacted screenshots
- Credential separation
- Audit logs
- Vendor review
- Incident process
Use aggregated, pseudonymized, or de-identified reporting where possible. Follow platform terms and applicable law.
100-Point Analytics Audit
| Area | Points |
|---|---|
| Metric dictionary | 10 |
| Native definition preservation | 10 |
| Inventory and listing data | 10 |
| Funnel measurement | 10 |
| Financial contribution | 15 |
| Advertising and attribution | 10 |
| Fulfilment/returns/service | 10 |
| Cohorts and segmentation | 10 |
| Reconciliation/data quality | 10 |
| Ownership/privacy | 5 |
Unreconciled financials or knowingly mixed incompatible definitions are critical failures.
Final Checklist
- [ ] Metric dictionary approved
- [ ] Native platform definitions retained
- [ ] Orders separated from units
- [ ] Visits separated from views/impressions
- [ ] Organic and paid data separated
- [ ] Contribution includes variable costs
- [ ] Inventory age and value tracked
- [ ] Seller quality monitored natively
- [ ] Returns have root-cause categories
- [ ] Median and 90th percentile used for timing
- [ ] Cohorts include sold and retired inventory
- [ ] Currency and time zones documented
- [ ] Raw exports preserved
- [ ] Orders, fees, refunds and payouts reconciled
- [ ] Data quality visible
- [ ] Each KPI has owner and action
- [ ] Buyer data minimized and protected
Frequently Asked Questions
What is the most important marketplace KPI?
No single KPI is sufficient. Contribution, inventory accuracy, account health and customer outcome are foundational; funnel metrics explain where improvement is possible.
Is revenue the same as profit?
No. Revenue does not subtract product cost, fees, advertising, shipping, returns, labour or fixed operating costs.
Can I compare conversion rates across Etsy and eBay?
Only after understanding each platform's numerator, denominator, attribution and traffic definition. Preserve native rates and avoid false precision.
What is a good conversion rate?
It depends on channel, category, price, inventory type, traffic quality and definition. Use internal baselines and comparable cohorts.
How often should dashboards refresh?
Operational exceptions may need near-real-time or daily updates. Strategic and financial reporting can be weekly or monthly after reconciliation.
Should likes and favourites be treated as demand?
They are engagement signals. Validate them against orders, accepted price, inventory age and contribution.
How should advertising be measured?
Track native attributed results, complete contribution after ad cost, and incrementality where a valid test is possible.
Why do platform reports and payouts differ?
Timing, pending transactions, refunds, reserves, fees, adjustments, tax and multi-order payout grouping can create differences. Reconcile transaction-level records.
Final Recommendation
Measure the marketplace in layers.
Keep native platform definitions intact. Normalize orders, SKUs, inventory and financials separately. Follow the funnel from eligibility to contribution. Reconcile money. Segment by product and cohort. Assign owners and actions.
A useful marketplace dashboard does not display the most numbers. It makes the next responsible decision obvious—and shows whether that decision improved customer outcomes and sustainable contribution.
Related posts
Canonical Tags: A Practical Guide for Business and E-Commerce Sites
A practical canonical tags guide for business and e-commerce websites covering duplicate URLs, rel canonical, redirects, sitemaps, hreflang, product variants and audit workflows.
Read article →
Cross-Listing Inventory Management Without Overselling
A practical cross-listing inventory architecture covering canonical SKUs, stock states, reservations, marketplace mappings, order events, synchronization, reconciliation and recovery.
Read article →
Depop Pricing Strategy: Fees, Offers and Profit
A practical Depop pricing system covering market research, location-based fees, buyer costs, offers, discounts, bundles, shipping, boosting, returns and contribution.
Read article →
Author
Anushka Dahanayake
Anushka Dahanayake is the founder of ANUSHKA DAHANAYAKE (PVT) LTD, building SEO-driven content, digital services, and revenue platforms for businesses in Sri Lanka and worldwide.