Budget for four specialists rarely arrives at once. A board signs off an ecommerce manager in the spring, agrees to revisit a developer in the autumn, and leaves the CRO specialist and the UX designer on a list that gets shorter with every planning round. The choice described in that meeting as building a team or retaining an agency is therefore usually a narrower one: which single capability arrives first, and who covers the other three while the store keeps trading.
Salary comparisons tend to dominate the conversation, and they are the least informative part of it. The more revealing question is how much genuine work each of the four roles would have in an average month, because that is what separates a business that should hire from one better served by an ecommerce agency or an equivalent external team.
What the Decision Actually Turns On
Hire internally when the work is continuous, high in volume, and dependent on knowledge that takes months to accumulate: the catalogue, the margins, the suppliers, the way customers in your category actually buy. Retain one external team when the business needs four partial capabilities at the same time and cannot justify, or cannot wait for, four appointments. Most established retailers settle somewhere between the two, with one internal owner and bought-in capacity behind them.
HubSpot’s guidance on when to outsource marketing work is usefully specific about the conditions. It recommends keeping activity in-house where the current technology is already generating strong returns and the business has a trained, experienced team with the capacity to handle both long-term campaign work and ad-hoc tasks. It suggests considering outsourcing where team members are “wearing too many marketing hats”, consistently missing deadlines, or are hampered more than helped by their tooling. The test is capacity and fit rather than principle, and it can point either way for two retailers of the same size.
Shopify describes the arrangement that most often settles the argument in practice, where a company does both: a generalist stays in-house to manage the agency relationship while the external team handles execution. That structure tends to survive because it assigns the two things that cannot sensibly be bought, prioritisation and accountability, to someone on the payroll.
Why Four Roles Are Rarely Four Full-Time Jobs
Each of the four roles has a different relationship to volume, and at the size of business where this decision usually arises, only one of them is reliably busy.
A CRO specialist’s output is bounded by traffic and by how many tests a store can run to a readable conclusion. A retailer with a few hundred conversions a month may only complete a handful of valid experiments in a quarter, and a full-time specialist will spend much of the intervening time on analysis, instrumentation and waiting. That is not wasted work, but it is rarely a full week of it. Where traffic supports continuous testing, the arithmetic reverses and a permanent specialist becomes the cheaper structure, because the alternative is paying external day rates for work that never stops.
UX design is lumpier still. It follows the roadmap rather than the calendar, which means concentrated effort around a redesign, a new category structure or a checkout change, separated by stretches with nothing substantial to build. Designers without a pipeline often fill the gap with visual refinement nobody asked for, which is one way stores acquire redesigns they cannot explain commercially.
The ecommerce manager is the role closest to genuinely full-time. Trading, merchandising, promotional planning, supplier coordination and the daily arbitration between marketing and operations do not pause. Shopify’s advice on building a marketing team points the same way, suggesting that businesses concentrate on essential roles such as a marketing manager in the early stages and add specialists as they grow.
A developer sits at either extreme depending on the backlog. Where there is a real programme of integration work, performance engineering or a platform migration, the role is full and then some. Where there is not, the queue fills with bug reports and small requests, and the business ends up with a capable engineer maintaining a ticket system. That pattern is common enough to have its own discussion here, in the article on who owns ecommerce growth when the developer only handles tickets.
Read together, these differences reframe the cost question. The comparison is not four salaries against a retainer. It is the fraction of each role the business will actually consume, against what those fractions cost when bought separately.
Signs the Workload Genuinely Supports Permanent Specialists
- Testing is limited by ideas rather than by traffic. If the store has enough conversions to read a result inside two or three weeks, and the backlog of evidenced hypotheses runs longer than the calendar, a permanent CRO specialist will be occupied.
- The roadmap is funded twelve months out. Design and development are only continuous when there is agreed, budgeted work waiting behind the current project, rather than an intention to keep improving things.
- Catalogue complexity is the main constraint. Deep product data, configurable items, multi-market pricing or B2B rules reward someone who carries that knowledge rather than rebuilding it at the start of each engagement.
- The store is the primary channel rather than one of several. Where ecommerce is a share of a wider business, specialist time tends to get split, and split roles are difficult to recruit for and easy to lose.
- There is someone senior to manage specialists. Four individual contributors without a manager will each optimise their own area competently while the roadmap between them stays unowned.
What a Ecommerce Agency Replaces, and What It Does Not
BigCommerce defines an ecommerce agency as “a specialized service provider that helps businesses build, optimize, and grow their online stores”, with services spanning design and user experience, development and integrations, platform migrations, conversion rate optimisation and search. That description maps almost exactly onto the four roles under discussion, which is why the comparison is a real one rather than a marketing construct.
What an external team genuinely replaces is breadth at partial utilisation, plus the months a business would otherwise spend recruiting. Shopify’s summary of when an ecommerce agency makes sense names three situations: an in-house team that has hit a ceiling on what it can do, specialised expertise the business cannot hire fast enough, and a growth phase requiring capacity and speed together. Buying delivery this way is ordinary rather than exceptional. Salesforce reports that 70% of its customer implementations are led by consulting partners, which indicates how implementation work is normally resourced in that ecosystem. It says nothing about whether any particular retailer should outsource, and it should not be read as evidence that external delivery performs better.
What an ecommerce agency does not replace is the internal judgement that decides what gets built. Commercial priorities, margin tolerance, supplier constraints and brand positioning stay inside the business whether or not anyone has been hired, and an external team working without that input will default to generic best practice. HubSpot raises a more concrete version of the same risk, noting that external agencies often retain control of critical assets, including contact databases, campaign data and platform accounts. That turns a later transition into a project of its own. Ownership of accounts, data and code is worth settling before the first invoice rather than during a handover.
Four internal hires, one ecommerce agency and the hybrid arrangement compare as follows across the criteria that tend to decide the outcome.
| Criterion | Four internal hires | One external team | Hybrid: internal owner plus external capacity |
| Best suited to | High-volume stores with continuous specialist work and complex internal knowledge | Businesses needing four partial capabilities at once, or a defined programme with an end date | Established retailers with one clear internal owner and an uneven workload |
| Time to full capability | Roughly six to twelve months across four recruitment cycles | Weeks, subject to scoping and onboarding | Weeks for capacity, months for the internal appointment |
| Cost structure | Fixed salaries, employer costs, tooling, recruitment fees | Retainer or project fees that vary with scope | One salary plus variable external spend |
| Utilisation risk | High where volume does not support four roles | Low, since the business pays for what is scoped | Moderate, concentrated in the internal role |
| Brand and catalogue knowledge | Strongest; accumulates and stays | Must be transferred, and re-transferred when account staff change | Held internally and briefed outward |
| Continuity risk | One resignation can remove an entire discipline | Contracts end and account teams get reassigned | Lowest of the three |
| Who sets priorities | The internal manager, where one exists | Whoever the client nominates; otherwise the agency decides by default | The internal owner, explicitly |
| Typical failure mode | Specialists underused and drifting into low-value work | Activity delivered competently without commercial ownership | The internal owner becomes a bottleneck |
No column wins on every row, which is the reason for setting them out together. Internal teams are strongest exactly where external ones are weakest, on accumulated knowledge and availability, and weakest where external ones are strongest, on breadth and on absorbing an uneven workload. The hybrid column reads like a compromise and usually behaves like one: somewhat more expensive than a pure agency arrangement, and considerably more resilient than four unmanaged specialists.
Continuity deserves more weight than it normally receives in these discussions. A four-person internal team contains four single points of failure, and losing the only person who understands the analytics implementation is a different order of problem from losing a supplier. An external team spreads that risk across its own staff but introduces a commercial version of it, since contracts end, account teams change, and the accumulated knowledge leaves with them unless documentation was written into the deliverables.
Key takeaway: the question is rarely which model is better in the abstract. It is how many months of genuine specialist work the store can put in front of each role, and who inside the business will hold the priorities either way.
The Order of Arrival Usually Decides the Outcome
Because the four capabilities almost never arrive together, the sequence matters more than the eventual organisation chart. Three patterns cover most situations, and the right one depends on which constraint is currently binding.
Appoint the ecommerce manager first when the constraint is prioritisation rather than execution. Stores in this position usually have agencies, freelancers or a developer already doing competent work in isolation, with nobody deciding what matters most this quarter. Adding a fourth supplier will not resolve that, and adding a specialist may make the coordination problem slightly worse. The related comparison of a CRO specialist against an external ecommerce team works through the same trade-off for a single discipline.
Bring in a developer first when platform constraints are blocking everything else. Some of what a designer proposes may not be buildable on the current plan, and discovering that late is expensive in both money and credibility. Shopify’s documentation states that checkout UI extensions rendering on the information, shipping and payment steps are available only to stores on a Shopify Plus plan, so a checkout redesign on a standard plan meets a hard limit that no amount of design work removes. Someone needs to know where those limits sit before the roadmap is agreed, not after it has been presented to the board.
Engage a ecommerce agency first when the business has a roadmap but little evidence behind it. Analytics repair, behavioural research, a conversion audit and a first testing cycle are bounded pieces of work with a defined end, which makes them well suited to an external engagement and a weak reason to create a permanent post. Shopify makes a similar point about interim needs, noting that freelancers or an agency can fill a temporary position and offer a cost-effective way to establish whether a full-time hire is warranted at all.
How to Test Both Models Before Committing to Either
The comparison becomes much easier once it is grounded in a specific quarter rather than a general capability wish list. The following sequence takes a few weeks and tends to produce a decision that survives contact with the budget.
- Write down one quarter of real work. Not aspirations: the specific tests, design changes, integrations and analyses the business would fund in the next three months, with the name of the person who would approve each one. Roles are easy to justify in the abstract and much harder to justify against a finite list.
- Check what the platform permits before pricing anything. Plan restrictions, app conflicts and integration limits determine how much of that quarter is actually buildable. A capability that requires a plan upgrade belongs in the budget comparison rather than in a later surprise.
- Price both models across the same twelve months. Internal costs include employer contributions, recruitment fees, tooling, equipment and the management time four reports absorb. External costs are easier to read but vary widely: Shopify publishes indicative ranges of roughly $5,000 to $10,000 for basic store builds and $15,000 to $35,000 for mid-tier custom builds, and cites a Databox survey of more than 40 agencies in which most clients paid less than $30,000 a year. Treat published ranges as orientation, never as a quote for your own scope.
- Settle decision rights and asset ownership in writing. Who approves a test, who can stop a release, and whose name sits on the analytics property, the advertising accounts and the repository. This is the clause that determines how painful a change of supplier will be in two years, and it costs nothing to get right at the start.
- Start with a bounded engagement that has an end date. BigCommerce’s criteria for evaluating agencies apply directly before committing to either four appointments or an ecommerce agency retainer: look for clear pricing models, defined deliverables and realistic timelines, and treat vague proposals without clear success metrics, delayed responses and aggressive upselling as reasons to keep looking.
- Agree a review point and what would change the decision. A quarterly review against named metrics converts an open-ended retainer into a series of deliberate decisions. It also gives an internal appointment a fair comparison, which becomes difficult once four people are already on the payroll.
What Changes When One Team Covers Strategy, UX, Analytics and Development
The argument for a single external team is not primarily about price. It is that some problems only become visible when the disciplines sit close enough together to notice that they are looking at the same thing.
WD Market’s work with Profcentrs.lv is a reasonable illustration. The case study lists the engagement as four services: ERP integration and inventory synchronisation, a WooCommerce platform rebuild, brand consolidation strategy, and UX and conversion flow optimisation. The retailer was running two brands with stock that did not match between physical and digital stores, so shoppers were meeting out-of-stock messages on products that were sitting on a shelf somewhere. The published results include “Zero inventory mismatches” through automated real-time ERP synchronisation, and a returning customer rate that “increased by 71%”.
The part that matters for this decision is the diagnosis rather than the figures. A UX specialist looking only at the storefront would have found a usability problem and proposed storefront fixes. A developer looking only at the ERP would have found a synchronisation problem and proposed a data fix. They were the same problem, and a store can spend a year improving a product page that is faithfully reporting the wrong stock data. Four separate owners will each produce a correct answer within their own boundary. Whether anyone owns the boundary itself is the question worth putting to both models before choosing one.
Internal expert input required: add the first-90-days sequence WD Market typically uses when taking over from a mixed internal and freelance setup, and describe how account, analytics and repository ownership is transferred during that period.
Mistakes That Make Either Model Fail
Most of the disappointing outcomes in this area come from a small number of repeated errors, and none of them is really about which model was chosen.
- Comparing a salary directly to a retainer. A salary buys availability rather than output, and a retainer buys scoped output rather than availability. Setting the two headline figures side by side without stating what each one delivers usually flatters whichever option the reader already preferred.
- Hiring specialists without a manager. Four individual contributors reporting to a founder who has several other responsibilities will each do reasonable work in their own lane. The roadmap between them stays unowned, which is frequently the exact problem the hiring was meant to solve.
- Choosing an ecommerce agency on hourly rate. Rate is a poor predictor of cost per outcome. A team that already knows the platform may deliver in a third of the hours, so the lower rate can produce the larger invoice and the slower roadmap.
- Leaving accounts and code in someone else’s name. Analytics properties, advertising accounts, repositories and DNS records should sit with the business from the first week. Recovering them later is usually possible but slow, and it tends to become urgent precisely when the relationship is already strained.
- Buying execution when the real problem is decision-making. More capacity applied to an unagreed set of priorities produces more activity and a similar commercial result. Where nobody is deciding, neither four appointments nor an external retainer resolves it without someone being given the authority to choose.
Matching the Structure to the Work You Actually Have
The choice between four internal appointments and one external team is decided by the work sitting in front of the store, not by a preference for either model. Continuous, high-volume specialist work that depends on deep catalogue and customer knowledge argues for permanent roles. An uneven workload, a capability gap that cannot wait out a recruitment cycle, or a defined programme with an end date argues for bought-in capacity, whether that comes from an ecommerce agency or an independent specialist team.
One internal owner with external specialists behind them is where most established retailers end up, because it keeps prioritisation and accountability inside the business while the workload stays variable. Whichever direction the decision takes, the order in which capabilities arrive and the ownership of accounts, data and priorities deserve more attention than the salary comparison that usually opens the discussion.
From Team Structure to a Working Growth Roadmap
If the decision is still open, the more useful next step is usually an assessment of what a quarter of real work would contain and whether the store’s current data can support it, rather than a hiring plan drafted in advance of the evidence. Apply for ecommerce growth support to get a review of analytics reliability, behavioural evidence and platform constraints, followed by a prioritised roadmap showing which work needs a permanent owner and which is better bought externally. Where the question runs wider than conversion, ecommerce strategy consulting addresses the same split at roadmap level.
To talk through how the roles and the workload would divide in your business, get in touch with the team and describe the quarter you are planning. Shorter notes on ecommerce team structure and conversion work are published on WD Market’s LinkedIn page.
Frequently Asked Questions
Is one external team cheaper than hiring four ecommerce specialists?
Usually on paper, though the comparison misleads unless both sides are priced over the same period. Internal costs extend well beyond salary to employer contributions, recruitment, tooling and the management time four reports absorb. External costs are scoped rather than fixed, so they fall when work pauses and rise when it accelerates. The figure worth calculating is cost per delivered outcome over twelve months. A store with enough continuous work to keep four people genuinely occupied will often find permanent appointments cheaper across several years.
Which role should an established store hire first?
In most cases the ecommerce manager, because the binding constraint is more often that nobody is deciding what matters this quarter than that nothing is being executed. A developer makes a stronger first appointment where platform limitations are blocking the roadmap and the business needs somebody who knows where those limits sit. Conversion and UX work is frequently better bought as a defined engagement first, since it produces the evidence that shows whether a permanent role is justified.
How long does it take to build a four-person ecommerce team?
Recruitment cycles rarely run in parallel, so a realistic expectation is several months per role including notice periods, and somewhere between half a year and a year before all four are in place and productive. Budget approval usually arrives in stages as well, which stretches the timeline further. The practical consequence is a long interim period with some capabilities missing, and that gap is what most businesses fill with freelancers or an external team while recruitment continues.
What should be agreed with an ecommerce agency before signing?
Ownership of analytics properties, advertising accounts, repositories and domains; who approves work and who can stop it; the specific deliverables for the first engagement with dates attached; the reporting format and cadence; and the notice period, including what happens to documentation when the contract ends. Since agencies frequently end up holding those assets by default, the ownership clause tends to matter more in the long run than most of the commercial terms negotiated alongside it.
Can a hybrid model work with a very small internal team?
Often yes, and it is the most common stable arrangement among established retailers. One internal person who understands the catalogue, the margins and the customer holds the priorities and manages the relationship, while specialist execution is bought as needed. The main risk is that the single internal owner becomes a bottleneck, particularly where they also carry trading and merchandising duties. Protecting part of their week for roadmap and review work is usually what keeps the structure functioning.
How do we know whether a full-time CRO specialist is justified?
Count how many tests the store could run to a readable conclusion in a quarter at its current conversion volume, then count how many hypotheses the business has real evidence for. Where the second number comfortably exceeds the first, a permanent specialist will have enough to do. Where the store can only complete a few valid experiments in three months, the same budget often buys more when spent on research and a smaller number of carefully designed tests.