A CEO can now obtain a pricing, cash-flow or marketing framework with a single question. If similar advice is widely available, where does Finance still create value? The hard part is not generating the framework. It is applying it to the specific business and turning it into a decision that can be tested and reviewed.
From evidence to a decision that can be reviewed
| Step | Focus |
|---|---|
| 01 Diagnose | What is actually happening? |
| 02 Choose | Which trade-off is fundable? |
| 03 Review | What changed, and why? |
Record assumptions, owners and review dates. Evaluate both the decision process and the outcome.
01A Reasonable Answer Can Start with the Wrong Problem
When the marketing ratio rises, the instinct is often to cut weak campaigns. The real cause may be a key SKU stock-out, uncompetitive pricing, a misleading listing, or sales figures not adjusted for returns. Each diagnosis requires a different response.
Finance should trace the numbers and speak with operators before recommending action. A sophisticated model cannot fix a wrong diagnosis.
02Decision-Ready Information Does Not Appear Automatically
Sales may report GMV, Marketing may report attributed revenue, and Accounting may report revenue after returns. SKU reports often omit return costs or affiliate fees. These differences are common and material.
Finance adds value by aligning the bases, checking reliability, and explaining what the data does — and does not — show. AI can help reconcile and flag exceptions. Someone still has to define what the numbers represent and what is missing.
03Hard Decisions Involve Competing, Reasonable Goals
Marketing needs time and spend to acquire customers. Operations needs stock to protect sales. The CEO wants speed. Finance must protect liquidity. All of these objectives can be legitimate at the same time, yet they are rarely fully fundable together.
Good decisions make the trade-offs explicit: expected contribution, capital required, acceptable downside, and what is being deferred. Sound judgment does not guarantee a good outcome. It requires transparent assumptions and the ability to adjust.
04The Same Data Can Still Call for a Test
Suppose segment A shows high margins but weak repeat purchase, while segment B shows lower margins but stronger repeat purchase. Historical averages alone do not prove that B should be scaled. New customers may behave differently, or new offers may compress margins.
Finance can design a limited test, cap the capital at risk, and measure cumulative cohort contribution. Structured learning reduces uncertainty before larger commitments are made.
05A Decision Must Move Through the Organisation
A price increase may conflict with distributor agreements. Lower inventory targets may conflict with supplier MOQs. A supplier change may affect quality. Commercial Finance needs to understand these constraints, negotiate practical alternatives, and agree clear ownership.
After approval, track both actions and results. When value does not appear, separate flawed assumptions from incomplete execution. Reporting variance alone is not enough.
06Use AI to Free Time for the Harder Work
AI should handle aggregation, modelling, data checks and initial hypotheses. That creates capacity for the work that still requires judgment: understanding customers and operations, framing real choices, and following decisions through.
Before any recommendation, two questions remain useful:
- What about this business changes the generic answer?
- What evidence would cause us to change the decision?
Key takeaway
Finance adds value through context-sensitive diagnosis, explicit trade-offs, testable decisions, and disciplined follow-through until results are clear enough to learn from.
All numerical examples are hypothetical, not client results. Definitions, assumptions and business context affect the conclusion.
Back to E-commerce ↗