Ecommerce marketing agencies sit closer to revenue than almost any other agency type. That makes evaluation easier in one way — the numbers are visible — and harder in another, because a lot of what moves those numbers is not the agency's work.
What an ecommerce agency should own
Most direct-to-consumer engagements cover some mix of paid acquisition, lifecycle email and SMS, merchandising and product page conversion, and reporting that ties spend to contribution margin rather than platform-reported revenue. Very few agencies are strong at all of it. Decide before you brief which part is your actual constraint.
- Acquisition problem: not enough qualified traffic at a workable cost. You need a [paid media](/agencies/paid-media) or [growth marketing](/agencies/growth-marketing) partner.
- Conversion problem: traffic arrives and does not buy. You need [conversion optimization](/agencies/conversion-optimization) and [UX/UI](/agencies/ux-ui) work.
- Retention problem: first orders do not become second orders. You need [email and retention](/agencies/email-retention) capability.
- Platform problem: the store itself limits you. You need [Shopify](/agencies/shopify) or [web design](/agencies/web-design) build skills.
Questions that separate operators from generalists
- Which metric will you be accountable for, and how is it calculated? Ask them to define it in writing. Platform-attributed ROAS, blended MER and contribution margin tell very different stories about the same month.
- How do you handle attribution after iOS privacy changes? An honest answer describes triangulation between platform data, a measurement baseline and incrementality testing — not a claim of perfect tracking.
- What is your testing cadence, and what happens to a losing test? Look for a documented backlog, not ad-hoc creative refreshes.
- Who touches the account day to day? Ask for names, hours and seniority, then check that the people in the pitch appear in the plan.
- What will you need from us? Agencies that work well in ecommerce are specific: product feed access, margin data, a creative shooting cadence, a customer service feedback loop.
Reading their evidence
Ask for two or three engagements in your revenue band and category, with the starting position, what changed, over what period, and what else was happening in the business at the time. A result quoted without a baseline or a timeframe is not evidence. Where an agency cannot share client names, an anonymised account with real numbers and a named reference call is a fair substitute.
Be sceptical of percentage growth figures on small bases, results reported during a period that included a funding round or a major PR moment, and screenshots of platform dashboards with no context.
Fee models and what they do to behaviour
- Flat monthly retainer: predictable, and the agency has no incentive to inflate spend. Watch for scope drift.
- Percentage of ad spend: aligns the agency with bigger budgets, not better margins. If you use it, cap it or taper the percentage as spend rises.
- Performance component: workable when the metric is one both parties control and can audit. Avoid tying a bonus to a number the agency also reports on unaudited.
Structuring the first engagement
A 60 to 90 day initial scope with a defined diagnostic phase protects both sides. Expect an audit, a measurement baseline agreed in writing, a prioritised roadmap, and a first wave of tests. Set a review date before you sign, and agree what evidence will be on the table at that review.
Warning signs
- Guaranteed ROAS or revenue figures before any access to your data.
- No question about your margins or your fulfilment costs.
- A single named strategist across a large roster of accounts.
- Reporting that never shows a losing test.
Next steps
Browse [ecommerce marketing agencies](/agencies/ecommerce-marketing) in our directory, or read our [methodology](/methodology) to see how we score the agencies we have researched. Every claim on an agency profile is labelled with its source and verification status.