Two supplier arrangements can look identical from the outside. The same monthly invoice, the same weekly call, often the same competent people doing the work. The difference sits in a smaller place: where the next piece of work comes from. In one, the retailer decides what should happen and the supplier prices it. In the other, the supplier arrives with a view on what the store should do next, and answers for that view when it does not survive contact with the data.
An ecommerce growth partner is not a higher grade of vendor. It is a different allocation of one responsibility: who owns the question of what the store works on next, and who carries the consequences of getting that order wrong. Both models are legitimate. The wrong one tends to be chosen for reasons that have little to do with how good the supplier is.
What an Ecommerce Growth Partner Does That a Vendor Does Not
A vendor relationship is defined by the brief. The retailer specifies the work, the supplier prices and delivers it, and success means the specification was met on time and within budget. That is a demanding discipline in its own right, and a business that knows precisely what it wants built is usually better served by paying for execution than by paying somebody to think on its behalf.
An ecommerce growth partner takes one item across the table: the prioritised list itself. The supplier produces the diagnosis, argues for a sequence of work, and is judged on whether that sequence improved the commercial result rather than on whether each ticket closed. In exchange for influence over what gets built, it inherits the risk of choosing badly.
There is a test that can be applied to an existing relationship without renegotiating anything. Ask who wrote the current quarter’s list of work, and who would be uncomfortable if that list turned out to have been the wrong one. If both answers point inside the business, the arrangement is a vendor relationship whatever the contract calls it. That is not a criticism of either side, only a description of where the thinking happens. The underlying issue of who owns ecommerce growth when a developer only handles tickets sits beneath most of these arrangements.
Five Differences That Show Up in Practice
The distinction between a vendor and an ecommerce growth partner becomes concrete in five places. Each can be observed in a current relationship before any decision is made.
1. Where the next piece of work comes from
In a vendor arrangement, work enters through a request, so the queue reflects whoever asked most recently and most insistently. Over time that favours visible problems over expensive ones. A partner proposes the queue and defends it, which means being able to say that a requested change is not worth doing this quarter. A willingness to decline politely, with reasoning attached, is the clearest signal that responsibility has moved rather than been renamed.
2. What the supplier is measured on
Delivery metrics suit vendor work: tickets closed, releases shipped, hours used against estimate. They rarely tell a business whether the store improved. An ecommerce growth partner needs measures that still mean something in a quarter with few releases: conversion by device and entry channel, revenue per session, and customer lifetime value, described by Klaviyo as “a metric that projects the amount of money a customer will spend with your company over the entire time that they do business with you”. Agreeing those measures before the work starts is what prevents an argument about them a year later.
3. What happens when a change does not work
A share of carefully reasoned changes will not produce the effect expected of them. That is ordinary, and it is where the two models diverge most sharply. A vendor has met its obligation once the change is live and correct, so a disappointing result becomes a fresh request. A partner should have said in advance what would count as failure, measured it, and folded the answer into the next decision. Without that loop, a store accumulates completed work rather than understanding of its own customers.
4. How much context the supplier is expected to hold
Vendor work can be completed from the brief alone. Partnership work depends on knowing how customers behave, which is slower and harder to bill for. Microsoft Clarity’s own product team describes recordings as “a really easy way to identify the reason a metric dropped without having to involve an analyst/engineer”, which is close to the standard worth holding a partner to. Somebody on the supplier side should be able to explain why a number moved without waiting for the next reporting cycle.
5. What the agreement actually commits to
Most supplier contracts describe hours, rates and response times. HubSpot describes a service level agreement as “a contractual document that establishes specific, measurable standards between a service provider and customer”, and for a growth engagement the standards worth writing down are rarely about response time. They concern what gets researched, how priorities are set, what evidence must accompany a recommendation, and who decides when the two sides disagree. An arrangement that never recorded those tends to drift back into a ticket queue without anybody deciding that it should.
Signs the Current Relationship Has Reached Its Limit
The move between models is rarely triggered by poor delivery. More often it follows an accumulation of small signals that nobody is holding the whole picture.
- Requests are completed accurately, yet commercial numbers have been flat for several reporting periods, which points at what is being requested rather than at how it is built.
- Nobody can name the store’s three biggest conversion problems without opening a tool first, which usually means no one has been asked to hold that view.
- Week-long work is repeatedly displaced by afternoon work, because the afternoon jobs have named requesters and the longer ones do not.
- Analytics, heatmap and testing tools are installed, but no single person is accountable for reading them.
- Each new supplier begins by asking for context a previous one had already assembled, a sign the knowledge lives with the vendors rather than the business.
None of these on its own justifies changing supplier. Together they describe a business paying competently for execution while the more valuable work, deciding what deserves that capacity, goes undone.
Comparing the Three Realistic Arrangements
The choice is rarely binary. Most established retailers are choosing between three arrangements, and the third is the least discussed.
| Dimension | Vendor or supplier list | Ecommerce growth partner | Internal lead with vendors |
| Who sets priorities | The retailer, request by request | The supplier, with the retailer approving or refusing | A named internal owner, informed by suppliers |
| What the retailer is buying | Capacity and specialist skill | Diagnosis, sequence and execution | Capacity, plus one salary for the judgment |
| Usual measure of success | Delivered to specification | Movement in commercial measures agreed in advance | Both, split by function |
| Internal time required | High: somebody must brief, sequence and chase | Moderate: decisions, access and review | Highest at first, lower once the role is established |
| Speed to first useful output | Fast, once a brief exists | Slower, because research precedes recommendations | Slowest, because recruitment comes first |
| Main risk | The request queue quietly becomes the strategy | Dependence on one supplier’s judgment | A single point of failure in one person |
| Cost pattern | Variable, tracking the volume of requests | Retainer, less sensitive to volume | Fixed salary plus variable supplier cost |
| Suits a business that | Knows what it wants built | Has no internal owner of growth | Has enough continuous work to justify the role |
Two rows deserve more than a table cell. Internal time is routinely underestimated in vendor arrangements: a brief good enough to price is not the same as a brief good enough to be worth building, and the gap is senior time nobody budgeted for. The risk row runs the other way. An ecommerce growth partner concentrates judgment in one supplier, which is efficient until that judgment is wrong, and the protection is not a second supplier but a written record of why each decision was taken.
Scope is the other thing a longer arrangement tends to change. The RIPO International case study lists work spanning a full UX and UI redesign, consolidation of separate country sites into one platform, checkout flow optimisation, and CRO strategy and implementation, and records “+150% Revenue growth €200K → €500K annual turnover” for that programme as a whole. The figure belongs to the combination of work rather than to any single change, and what is available to another store depends on its own starting position.
The hybrid is worth attention because it is where many retailers settle once the first model stops working. Where the open question is whether to build that capability internally, the comparison between a single CRO hire and an external team covers the staffing side in more detail.
When a Vendor Relationship Is the Better Choice
Paying a supplier to set priorities is wasteful when the priorities are already settled, and several situations qualify.
A replatforming project, an integration with a fixed specification, or a compliance deadline are defined by external constraints rather than by diagnosis. A vendor priced on delivery is usually cheaper and safer for that work, and asking the same supplier to own growth strategy alongside it often produces a slower project and a weaker opinion on both.
The same reasoning applies when the business already employs somebody who does the thinking. If an ecommerce manager or head of digital holds the roadmap, reads behavioural data and can defend a sequence internally, an ecommerce growth partner duplicates a role that is already filled. The more useful purchase is capacity: development, design or research bought against a plan that exists.
Order volume is the third condition, and often the decisive one. Below a certain level of traffic, experiments take too long to settle disagreements, so an ongoing retainer may struggle to demonstrate its own effect. A short diagnostic engagement followed by vendor-priced implementation tends to serve that business better, and an evidence-based usability review is a reasonable place to begin.
Key takeaway: a partnership exists when the supplier can be wrong in public about what the store should do next. If nobody on the supplier side carries that exposure, the arrangement is execution with a different label, and it should be priced and managed as execution.
How to Test a Supplier Before Committing to a Longer Agreement
A twelve-month agreement signed on the strength of a proposal is a bet on a document. A shorter piece of paid work reveals considerably more, and the sequence below is designed to produce evidence rather than reassurance.
- Buy a bounded diagnosis first. A fixed-price review of a defined area, with a deadline, shows how a supplier works before either side commits to a year of it.
- Give real access rather than a summary: analytics, behavioural tools, platform admin and, where possible, order data. A supplier working from a briefing document can usually only confirm what the business already suspects.
- Ask for the problems in priority order with the reasoning attached. The order matters more than the list: most reviewers can find twenty issues, and the useful output is which three deserve capacity and why the rest do not.
- Look for a recommendation that was ruled out. A review that rules nothing out has either not looked hard enough or is reluctant to disagree with the person paying for it.
- Check that findings are separated by cause. Usability problems, platform limitations, tracking faults and merchandising decisions need different owners and budgets, and a list that mixes them is difficult to act on.
- Agree the measures and the baseline before the first change ships. Baselines established afterwards are open to interpretation, and disputes about measurement are harder to settle than disputes about design.
A diagnosis concluding that the checkout is sound and the problem sits upstream on the product pages is worth more than one recommending a broad programme, and it is a fair filter on whether a longer arrangement is warranted at all.
What the Retailer Has to Supply for the Arrangement to Work
An ecommerce growth partner arrangement fails from the client side more often than it is given credit for, usually for practical reasons rather than commercial ones.
Access comes first. A partner needs administrative access to the platform, to analytics and to the behavioural tools, not periodic exports. Data history comes second, and it carries a hard limit. Google Analytics 4 offers user and event data retention of 2 months or 14 months on standard properties, and data that reaches the end of that period is deleted automatically each month. A store left on the shorter option may find the seasonal comparison it wanted is simply not available, which is better corrected before an engagement begins than during it.
Change control comes third. Both major platforms support testing before publishing. Shopify advises duplicating a theme because “before you customize a theme, it’s a good idea to make a backup copy so that you can discard your changes and start again if you need to”, while WooCommerce recommends a staging site, “a clone of your production site” where changes can be tested “without your live site being affected and potentially losing revenue”. A supplier shipping straight to production is not working faster. It is moving the risk into the trading week.
The last requirement is a decision-maker with enough standing to approve a sequence and decline a request from another department. Without one, the roadmap becomes another opinion competing with the marketing calendar, and the arrangement reverts to the model it was meant to replace.
Choosing the Arrangement That Fits Your Store
The decision is not really about supplier quality. It is about where responsibility for choosing the next piece of work should sit, and whether anyone inside the business has the time and the evidence to hold it. Where that person exists, vendors priced on delivery are usually the more economical arrangement. Where the role is vacant, the work does not disappear. It gets done by the request queue, in the order the requests happened to arrive.
Before changing anything, apply the test described earlier to the current arrangement: who wrote this quarter’s list, and who would be uncomfortable if it proved to be the wrong one. The answer usually indicates whether a bounded diagnosis or a longer engagement is the appropriate next step.
From Supplier Management to a Prioritised Plan
Where responsibility for priorities is currently unclaimed, the sensible first purchase is a bounded piece of diagnostic work rather than a long agreement. WD Market’s CRO and growth support is structured that way for established retailers: research first, then a prioritised sequence with the reasoning attached, then implementation measured against baselines agreed in advance. Businesses that need the commercial direction settled before any of that may be better served by ecommerce strategy work, and those unsure whether the current build can carry the plan at all by a technical audit.
To work out which arrangement suits your store, send the details through the contact page: the platform you run, roughly how many orders a month you process, and who decides today what gets built next. Shorter notes from this kind of work appear on WD Market’s LinkedIn page.
Frequently Asked Questions
Is an ecommerce growth partner simply an agency on a longer contract?
Contract length is a consequence, not the definition. Longer terms exist because research, implementation and measurement take time to complete a cycle. What distinguishes an ecommerce growth partner is that it proposes the work rather than receiving it, and accepts being judged on commercial movement rather than delivery alone. A twelve-month contract to complete whatever tickets arrive is still a vendor arrangement, priced less flexibly.
Can an existing supplier be turned into a growth partner?
Sometimes, and it is worth attempting before running a search. The supplier needs three things it may not currently have: access to behavioural and commercial data, permission to propose work nobody asked for, and a named person on the retailer’s side who can approve or refuse a sequence. Where a team has spent years responding to requests, the shift can be uncomfortable, so test it on one quarter of work rather than announcing it as policy.
How long before this kind of arrangement should show results?
Evidence of progress should appear well before commercial movement does. Within the first months, expect a trustworthy measurement setup, a prioritised set of problems with reasoning attached, and a few completed changes. Revenue effects depend heavily on order volume, because low traffic lengthens the time any result takes to become readable. A supplier that names a percentage before reading the data is offering an expectation, not a measurement.
Who should own the analytics accounts and the customer data?
The retailer. Accounts should be created under company ownership and suppliers added as users, rather than the reverse. The practical reason is continuity: when an arrangement ends, the business keeps its history, configuration and audience definitions. Retention settings belong to the account owner too, and a business that has never reviewed them may hold less history than it assumes.
We already have a marketing agency. Does this overlap?
There is usually less overlap than expected. Marketing work is generally measured on traffic, acquisition cost and campaign performance, while an ecommerce growth partner is concerned with what happens once a visitor reaches the store. Conflict tends to arise over measurement rather than scope, since both parties can claim the same revenue. Defining in advance which metrics belong to which relationship removes most of that friction.
What should we do if the supplier’s recommendations conflict with internal opinion?
Treat it as the arrangement working rather than failing. A supplier that never disagrees is providing execution, not judgment. Agree in advance how such disputes are settled: by testing where traffic allows it, by behavioural evidence where it does not, and by the named internal decision-maker where neither is conclusive. Recording the reasoning matters more than winning the argument, because it lets both sides review the decision later.