An ecommerce digital agency sells one of two quite different things, and most confusion in this purchase comes from not naming which. The first is build and platform work: designing and developing the store, migrating a catalogue, integrating payments, shipping and an inventory system, then handing it over. The second is growth: acquisition, merchandising, conversion rate work and retention on a store that already exists. Some firms do both competently, many claim both and are genuinely good at one. Buying a growth retainer from a build shop tends to produce a beautiful store with no traffic, and buying a build from a growth shop tends to produce a site that is slow in the ways that cost sales. This guide covers what each engagement contains, what moves the fee, and how to check the claims.
Build work versus growth retainers
Build engagements are fixed-scope projects with a defined end, and the risk is scope: catalogue migration, redirect mapping, tax and shipping logic, subscription or bundle rules, and the integrations to your enterprise resource planning or warehouse system are where budgets overrun. Ask for a written scope that names every integration and who owns each one, and insist the redirect map for old product and category URLs is a deliverable rather than an afterthought, because losing existing rankings at launch is the most expensive mistake in this category. Growth retainers are ongoing and the risk is different: an agency that owns paid media but not merchandising or site speed can only pull one lever. The most productive structure for an established store is a single team accountable for revenue with access to acquisition, on-site conversion and retention, since those three trade against one another constantly and splitting them across vendors turns every decision into a negotiation.
Product data is the part buyers underestimate
Ecommerce visibility depends heavily on structured product data being correct and complete. Google's ecommerce documentation is explicit that marking up product pages with structured data, including price, availability and review information where it exists, is what allows those pages to appear with rich results in search and to be eligible for shopping surfaces. Getting this right is unglamorous work: consistent identifiers, accurate availability that updates when stock changes, variant handling that does not create thousands of near-duplicate URLs, and canonical tags that point to the version you want indexed. Ask any candidate agency how they handle variants and faceted navigation, which is where most large catalogues generate crawl problems, and ask to see a client store's product markup rather than a slide about it. An agency that cannot explain your variant strategy in plain language will not fix your catalogue.
What moves the fee
Catalogue size and complexity move it most. A hundred simple products is a different job from ten thousand with variants, configurable bundles and regional pricing. Platform is the next input: work on a hosted platform is generally cheaper to maintain than a custom or headless build, which buys flexibility at the cost of needing developers permanently attached. Channel breadth is third, since running paid search, paid social, marketplaces and retail media is four disciplines rather than one. Then there is creative volume, which in paid social is the actual production line and often the largest line item on a growth retainer. Finally, whether media spend is billed flat or as a share of budget changes both the number and the incentives, so ask for the split explicitly. Fee structure and what it includes is the single most useful thing to compare across a shortlist, and it is worth putting every quote onto the same page before any of them are discussed.
How to vet a shortlist
Shop the agency's clients. Buy something small from two of their stores on a phone, note how fast the pages load, how the search and filtering behave, how many steps checkout takes and what arrives in your inbox afterwards. Twenty minutes of that tells you more than any case study, because it is the actual product of their work. Then ask for a client where revenue did not grow and what happened, and for the reason a recent engagement ended. Ask which of the store's numbers the agency will be accountable for: revenue, contribution margin after media, conversion rate, or something softer. Insist on access to your own accounts and platform in your own name, and on owning creative files and product data exports. Finally, ask what they would stop doing in month one. A candidate who only adds work has not looked closely at what you already run.
Questions people ask about ecommerce digital agency
Should I use one agency for build and growth?
One team is simpler and avoids the blame gap when a launch hurts performance, but only if the firm genuinely does both. Ask which of the two is the majority of their revenue and to meet the leads for each discipline. If the answer is vague, use a specialist for the build and appoint the growth partner early enough to review it before launch.
How do I protect rankings during a replatform?
Treat the redirect map as a paid deliverable, freeze URL changes that are not necessary, keep page content and internal linking intact where you can, and benchmark your top landing pages before the switch. Plan for a dip and monitor daily for the first fortnight so that a broken template is caught in days rather than at the next monthly report.
What should a growth retainer report on?
Revenue and contribution after media cost, blended acquisition cost against contribution margin, conversion rate by device, and repeat purchase rate. Channel-level return on ad spend belongs in the detail, not the headline, because it is the number most easily improved by shifting budget onto customers you would have kept anyway.
Is conversion rate work better value than more traffic?
Often, on a store with meaningful existing traffic, because the gain applies to every visitor you already pay for. It is not a substitute for acquisition on a low-traffic store, where there is not enough volume to test with. Ask a candidate to say which of the two they would fund first and to justify it from your own numbers.