Omnichannel: measuring transfers, service and margin.
Key distinctions
Three conditions before calculating.
Origin and transaction separated
The channel that collects the payment is not always the one that creates the demand. Documenting origin, assistance and transaction avoids rewarding the last touchpoint at the expense of the journey.
Full cost to serve
Picking, stock transfer, store compensation, waiting, return and customer contact are assigned to the journey. Coordination costs can cancel a saving on delivery.
Counterfactual transfer
An omnichannel order is classed as new, transferred from a channel, brought forward in time or won from a competitor. Only the first and last categories clearly create volume.
Method
Assessing an omnichannel journey without double attribution
- 01Trace the steps
Record search, advice, basket, payment, picking, collection, return and service with timestamp, place and customer identifier where legitimate.
- 02Assign the costs
Measure cost per step, failure rate, delay, substitution and store mobilisation; separate variable costs, capacity steps and investments.
- 03Estimate the transfers
Use geographic experimentation, progressive rollout or matched comparison to estimate what would have happened without the journey.
- 04Arbitrate the service
Compare contribution, satisfaction, delay, stock and repeat purchase. An immediate saving does not justify a collection promise that is frequently not kept.
The economic account of a click-and-collect journey
Value is tracked from acquisition to return. Each step carries an owner, a triggered cost and a risk of double counting.
Acquisition
Digital
−€20kTransaction
Web
+€72kPicking
Store
−€12kCollection
Store
service 97%Return
Cross-channel
−€9kTransfer
Retail displaced
−€16.8k2,000 collections produce €27,000 before cannibalisation
The journey generates €180,000 of net revenue and €72,000 of product margin. Picking, payment, returns and acquisition cost €12,000, €4,000, €9,000 and €20,000.
Direct contribution: 72,000 − 12,000 − 4,000 − 9,000 − 20,000 = €27,000. If 600 orders replace store sales that would have produced €28 each, the transfer is worth €16,800 and the genuinely additional gain falls to €10,200.
The journey stays positive but its profitability depends on the transfer. The rollout must measure counterfactual store sales and picking capacity.
Acceptance conditions
What must be true to act.
- 01
Demand origin, assistance and transaction are distinguished.
- 02
Costs of picking, stock, compensation, return and service are assigned.
- 03
Transfers between channels are valued with the contribution given up.
- 04
Thresholds of delay, availability and failure are tracked alongside the margin.
Limits
What this analysis does not prove.
- Fragmented identities make the journey incomplete and can bias the transfers.
- Team learning effects change after launch.
- A pilot area can differ from large cities in stock and store density.
References
Works cited.
- Verhoef, P. C., Kannan, P. K. & Inman, J. J. (2015), “From Multi-Channel Retailing to Omni-Channel Retailing”, Journal of Retailing.
- Neslin, S. A. et al. (2006), “Challenges and Opportunities in Multichannel Customer Management”, Journal of Service Research.
- Pauwels, K. & Neslin, S. A. (2015), “Building with Bricks and Mortar”, Journal of Retailing.