Hiring an ecommerce digital marketing agency is one of the few purchases where the sales process tells you almost nothing about the product. Every pitch deck contains the same case studies (a 4x ROAS, a store that went from $80k to $400k a month), the same channel logos, and the same promise of a dedicated team. Six months later, roughly half of these relationships end badly, and the post-mortem is almost always the same: the agency was measured on activity, not on profit, and nobody noticed until the retainer had eaten a quarter's margin.
This guide is written from the buyer's side. It covers what agencies actually do, what the pricing models really cost, the questions that separate operators from account managers, and the small number of metrics worth writing into a contract.
What an ecommerce agency actually does (and what it doesn't)
"Full service" is a range, not a definition. In practice, most shops sit in one of three groups.
Channel specialists run one thing properly: paid search and shopping, paid social, email and SMS, or SEO. They tend to be small, technically strong, and cheaper per channel. If your store has one obvious growth lever, they are usually the best value.
Full-funnel generalists run three to five channels plus reporting and some creative. Useful when you have no in-house marketing at all and need someone to own the whole picture. The risk is that they are excellent at one channel and mediocre at the rest, and you pay specialist rates across the board.
Growth partners take on merchandising, pricing, lifecycle, and site conversion alongside media. They cost the most and only make sense above roughly $500k in monthly revenue, where a one-point improvement in conversion rate funds the fee on its own.
What almost no agency does, whatever the deck says, is fix a product-market problem, repair a broken margin structure, or make a slow site fast without your developers. If your gross margin is 22% and your average order value is $38, no amount of media buying rescues the unit economics. Work that out yourself first with a profit margin calculator, because it determines whether an agency can help you at all.
The four pricing models, and what each one rewards
Fee structure is the single strongest predictor of how an agency will behave, because it decides what they optimise for when nobody is watching.
Flat monthly retainer. Typically $3,000 to $15,000 a month for ecommerce work. Predictable, easy to budget, and neutral: the agency earns the same whether you spend $20k or $200k. The weakness is that it rewards low effort once the account is stable. Guard against this with a defined scope of deliverables, not just a headcount promise.
Percentage of ad spend. Usually 10% to 20%, sliding down as spend grows. Common and mostly harmless at scale, but it quietly rewards spending more rather than earning more. An agency on 15% of spend has no financial reason to tell you that your worst campaign should be switched off. If you use this model, cap the percentage and pair it with an efficiency floor.
Performance or commission. Often 5% to 12% of attributed revenue, sometimes with a small base. Sounds perfectly aligned, and is the model most likely to cause an argument, because it makes attribution a billing dispute. Agree the measurement platform, the attribution window, and whether branded search and returning customers count before you sign anything.
Hybrid. A modest base retainer plus a performance component above an agreed baseline. This is the structure we see working most often. The base covers real operating cost so the agency can staff the account properly; the upside is tied to growth beyond what the store would have done anyway.
Whatever the model, convert the fee into a cost per acquired customer and add it to your media cost. An agency charging $8,000 a month on $60,000 of spend that produces 900 new customers is adding about $8.90 to your customer acquisition cost. That number belongs next to your media CAC in every review, and you can size it quickly with a CAC calculator. The broader framing of what an acquired customer should cost is covered in our guide to marketing customer acquisition cost.
Six questions that expose a weak agency
Most vetting checklists ask about experience, team size, and tools. Those questions are easy to answer well. These are harder, and the answers are diagnostic.
1. "What is our break-even ROAS, and how did you calculate it?" Ask this in the first call, before they have your P&L. A strong agency will refuse to guess and will tell you exactly what they need: gross margin after COGS, shipping, payment fees, and returns. A weak one will quote a target ROAS of 3 or 4 as though it were a universal constant. Break-even ROAS is simply one divided by gross margin, and it is the number every media decision hangs from. Run yours through a ROAS calculator so you can check their arithmetic on the spot.
2. "Show me an account you lost, and tell me why." Everyone has one. The useful signal is whether they can describe the failure in mechanical terms (attribution changed, the client cut inventory, the category got expensive) rather than blaming the client's ambition or the algorithm.
3. "Who touches the account day to day, and what else are they on?" Pitch teams and delivery teams are frequently different people. Ask for the name and workload of the person who will actually be in your ad account. An analyst carrying twelve stores cannot know your catalogue.
4. "How do you handle new versus returning customer revenue?" Blended ROAS flatters any agency working on a store with a strong repeat base. Insist on seeing new-customer performance separately. If they cannot split it, they are not instrumented well enough to be trusted with growth budget.
5. "What would you turn off in month one?" Good operators arrive with a subtraction list: dead campaigns, overlapping audiences, a discount that is cannibalising full-price sales. An agency that only proposes additions is selling scope, not judgement.
6. "What do you need from us to succeed?" The honest answer includes work on your side. Product feed fixes, faster creative approvals, access to margin data, developer time for tracking. An agency that says they need nothing is either very senior or has not thought about it.
Judge the audit, not the pitch
Most agencies will run a free audit as part of the sale. Treat it as a paid work sample and read it critically.
A shallow audit lists problems that any tool surfaces automatically: missing meta descriptions, low Quality Scores, unlinked assets. A serious audit connects observations to money. It will say something like: 38% of shopping spend goes to products with below-average margin, and shifting that budget to the top two collections should lift contribution by roughly $14k a month at current conversion rates. That sentence requires them to have looked at your catalogue, your margins, and your conversion data together.
Two further tells. First, whether they distinguish site problems from media problems. Many stores that "need better ads" actually need a better product page, and an agency that pushes media anyway is optimising for its own retainer. Our guide to ecommerce conversion rate optimization covers the checks worth doing before you increase spend. Second, whether they mention organic at all. Paid-only agencies often ignore the compounding channel entirely, even when a store's category and collection pages are one technical pass away from real traffic, which is the ground covered in our Shopify SEO optimization guide.
The metrics that belong in the contract
Vague goals produce vague accountability. Four numbers are usually enough.
Contribution margin after media and fees. The only figure that tells you whether the relationship is working. Revenue and ROAS can both rise while this falls.
New customer acquisition cost, reported separately. With a ceiling, not a target. A ceiling is enforceable.
Customer lifetime value by acquisition cohort. Agencies that buy cheap, low-intent traffic look good for one quarter and bad for four. Tracking cohort value by month of acquisition catches it early, and an LTV calculator gives you a working baseline before the first cohort matures.
A defined reporting cadence with raw data access. You should have direct login access to every ad account, analytics property, and email platform, in your own name, from day one. Any agency that resists this is holding your data hostage, and it is the single most common reason a bad relationship drags on longer than it should.
Set a review point at 90 days with explicit criteria. Not "we'll see how it goes", but a written statement of what has to be true by then. Most ecommerce channels give a readable signal within a quarter, and paid search faster than that, as covered in our guide to PPC for ecommerce.
When you should not hire an agency
Three situations where the money is better spent elsewhere.
Your monthly media budget is under about $10,000. Agency fees at that level consume 30% to 50% of working spend. A competent freelancer or a part-time in-house hire will do better.
You have no reliable tracking. An agency will spend the first two months rebuilding your measurement, billed at agency rates, and half of their early reporting will be guesswork. Fix conversion tracking first.
Your margin is thin and your catalogue is undifferentiated. Media amplifies whatever the business already is. If a competitor sells the same product $6 cheaper with free shipping, more traffic makes the problem more expensive, not less.
FAQ
How much does an ecommerce digital marketing agency cost? For a store spending $30,000 to $100,000 a month on media, expect $4,000 to $12,000 a month in fees, or 10% to 15% of spend. Below $10,000 of monthly media, agency economics rarely work in your favour. Above roughly $250,000, negotiate a hybrid deal with a lower percentage and a performance component.
How long before an agency produces results? Paid search and paid social should show a directional signal in four to six weeks and a defensible trend by 90 days. Email and lifecycle work compounds over two to three months. SEO is a six to twelve month commitment, and any agency promising otherwise is either misleading you or planning something you would not approve of.
Should I hire one full-service agency or several specialists? One agency is simpler to manage and cheaper in coordination cost, which matters more than most buyers expect. Specialists win when a single channel dominates your revenue or when you already have someone in-house who can hold the strategy together across vendors.
What is the most common mistake buyers make? Judging the relationship on ROAS alone. ROAS is a media efficiency ratio, not a profit measure. A store can hit 5x ROAS on discounted, low-margin products and lose money on every order. Contribution margin after media and fees is the number that settles the question.