Agency leaders were asked in early 2026 which of their services they expected clients to bring in-house. Promethean Research’s 2026 Digital Agency Industry Report, drawn from a survey base of 1,452 agencies, found web development at the top of that list, and found that web design and graphic design were the two categories that moved most between the 2025 and 2026 waves.
That measures what suppliers are worried about, not how many companies have done it. The distinction matters throughout. But suppliers watching their pipeline are reasonable early indicators, and the question is being asked more often than it was.
If you are asking it, the useful framing is not whether an internal team is cheaper. It is which parts of the work reward continuity and which parts reward range. Those are different properties, and most companies need both.
The work that benefits from an internal team is the work that recurs. The work that benefits from an outside team is the work that arrives rarely and badly needs to be right.
In this article
The decision is not build versus buy
Most companies frame this as a single choice: hire someone, or keep paying a supplier. Framed that way it turns into a cost comparison, and cost comparisons on this decision are almost always wrong, because the two options are not buying the same thing.
A website generates two kinds of work. Continuous work is small, frequent and endless. Copy changes, new pages, a product added, a price updated, a broken form. Episodic work is infrequent, large and consequential. A rebuild, a migration, a repositioning, a checkout redesign.
These reward opposite things. Continuous work rewards someone who is always available and already holds the context. Episodic work rewards someone who has done this particular thing many times recently, which almost by definition is not the person who does your continuous work.
Before comparing anything, count them. Go through the last twelve months of requests and sort each one into continuous or episodic. Most companies find the split is more lopsided than they assumed, and the answer usually falls out of the tally without further argument.
What an internal team is good at
Speed on small things. The gap between wanting a change and seeing it live is where most of the frustration with suppliers lives. An internal person removes a scoping conversation, a queue and an invoice from a fifteen-minute job. That is a real gain and it is the main reason in-housing succeeds when it succeeds.
Accumulated context. Someone who sits in your sales meetings knows why the pricing page is worded the way it is. They know which customers complained about which thing. That knowledge does not transfer well in a brief, and it compounds over years.
Willingness to do unglamorous work. Suppliers price by the project and are structurally reluctant to spend a day on something small. Internal teams will.
What it is reliably bad at
Anything that happens once every four years. A rebuild is the obvious case. An internal team will do one, learn from it, and then not apply that learning again until the lessons have gone stale. A supplier doing this work continuously carries pattern recognition an internal team cannot accumulate.
Breadth. A website in commercial use needs strategy, copy, design, frontend, backend, performance, accessibility and analytics. Promethean found that production staff make up 62.7% of employees at the average agency, which is what it takes to cover that range as a business. One or two internal people cover perhaps a third of it well.
Surviving a departure. This is the failure mode that does the most damage and gets scoped the least. A single internal person is a single point of failure holding undocumented knowledge. When they leave, the site becomes something nobody understands, and the true cost of the arrangement arrives eighteen months after the decision that caused it.
The coverage problem
Ask what happens when your internal person is on holiday, ill, or has left. For most companies the honest answer is that the site freezes. If the site is commercially central, a two-week freeze during a campaign is a real cost that never appeared in the comparison.
The same question applied to a supplier has a different shape. A supplier can also lose the person who knows your account, which is why it is worth asking directly who else there has worked on your site. But a firm of any size has somebody else, and the knowledge sits in documentation because it has to.
One requirement is worth insisting on either way. Whoever does the work, the site should be documented well enough that a competent stranger could pick it up. The reaction to being asked for that is itself informative.
Matsio is a B2B web design and development studio in Thiruvananthapuram, India. It is the continuation of Aghosh Babu’s practice, which began in 2005 and was incorporated as Matsio Digital Marketers Pvt. Ltd. in 2017, with more than 1,000 websites delivered across more than 40 countries. A meaningful share of that work is with companies that have their own internal team and bring in the studio for the episodic work, which is the arrangement described below rather than an argument against hiring. 85% of the studio’s clients return for further work, measured across every engagement since 2005.
The maintenance nobody scopes
A website is not a finished object. It has dependencies that expire. Platform updates, plugin updates, certificate renewals, deprecated integrations, browsers that change behaviour, security patches that cannot wait.
This work is invisible until it is not done, at which point it becomes urgent and expensive. When companies compare an internal hire against a supplier retainer, this is the line most often left out of the internal column, because it is not what the person was hired to do and not what anyone imagines when they picture the role.
Price it explicitly on both sides. Ask a supplier what the retainer covers and excludes. Ask, of an internal hire, in whose week it sits. An hour or two a month is not much until it is nobody’s.
What the pricing data suggests
One finding in the Promethean report is worth reading carefully if you are running this comparison. Agency concern about in-housing was inversely correlated with hourly rates. Firms charging more were less worried about losing work to internal teams.
The reading that fits is that in-housing competes with commodity execution and does not compete with judgement. Where a supplier is selling hands, an internal hire is a direct substitute and usually a cheaper one. Where a supplier is selling a decision about what to build and why, the substitution does not work, because you are not hiring that capability at that price.
That gives you a diagnostic. If everything you currently buy from a supplier could be described as execution against a specification you wrote, in-housing will probably work. If you rely on them to tell you what the specification should be, it will not.
The same logic applies to the hire. A person capable of deciding what the site should do commands a different salary from a person capable of building what they are told to build, and the business case is usually written for the second while the expectations are written for the first.
The arrangement that usually holds
The split that survives contact with reality is an internal person owning the continuous work, and an outside team owning the episodic work and the standard the continuous work is held to.
This works because it puts each kind of work where it belongs. The internal person gets speed and context on daily changes. The outside team handles the rebuild, and is available for the question that starts “should we”.
It fails in one specific way, which is worth naming. If nobody owns the overall standard, the site degrades through accumulated small decisions, none of which was wrong on its own. Whoever holds that standard needs to be named, and it needs to be somebody’s actual job rather than an assumption.
It also requires both sides to work in the same codebase without stepping on each other. That is a practical question about access, environments and review, and it is worth settling in the first week rather than the first incident.
When going fully internal is right
There is a threshold, and it is higher than most companies estimate. Full in-housing works when the website is the product rather than a channel, when changes are needed daily rather than monthly, and when you can hire at least three people so that no single departure stops the work.
Below that, you are not building an internal team. You are hiring one person and hoping. That can work, and sometimes does, but it should be understood as the bet it is rather than as the safe option.
The industry structure is a useful reference point. Promethean counts more than 71,000 agencies in North America, 87% of them employing fewer than fifty people. Firms that do this work full time mostly operate at small scale themselves, which is a fair indication of how much headcount it takes.
Being honest about what this data shows
The Promethean figures describe what agency leaders expect, gathered from a self-reported survey of suppliers. They are not a measurement of how many companies brought work in-house, or of how those decisions turned out. Nobody publishes that, and anyone claiming a precise figure for it should be asked where it came from.
What the data supports is narrower and still useful. The question is being asked more often, it is being asked about design as well as development, and suppliers whose work is hardest to specify are the least worried about it.
Related reading. How to choose a web design company in Thiruvananthapuram covers the evaluation criteria to apply if you decide to keep buying, and How choice research explains shorter service pages covers why a shorter list of offerings usually signals a clearer one.
The short version
Split the work before you compare the cost. Continuous work rewards an internal person who is always there. Episodic work rewards an outside team that has done it recently and often. Most companies need both, and the common arrangement is one internal owner for the daily work with an outside team for rebuilds and for the standard.
Before deciding, price the maintenance nobody scopes and answer what happens when your one person leaves. If those two answers are comfortable, in-housing will probably hold. If they are not, you have found the reason it usually does not.
A small thing and a big thing
One small thing to fix on your website today, and one big thing to learn that gets you more leads.
Your information is
safe with us. No spam.