I had a technical discussion with another web agency last week. They were advising a client to set up dedicated servers for a marketing website. Static pages, no user accounts, no dynamic content. Just a corporate site with a blog.
You do not need a server for that. A CDN handles it with faster delivery, lower cost, and virtually zero maintenance. The client would never know the difference, because the client is not a systems architect. They trust the person sitting across from them.
That conversation stuck with me. Not because of the technical disagreement, but because of what it reveals about how buying decisions work in this industry.
An industry without a filter
Web development is one of the few professional fields with no formal barrier to entry. There is no license, no board exam, no apprenticeship, and no mandatory certification. Anyone can register a business, build a website, and start selling advice.
Compare that to almost any other profession where you pay someone for their expertise. Architects need seven years of education and a licensing exam. Electricians go through an apprenticeship. Accountants pass certification exams. Even a plumber needs credentials before they touch your pipes.
Web developers need nothing. A laptop and a domain name is enough to start an agency.
The Stack Overflow Developer Survey 2024 found that 82% of developers learned to code through online resources rather than formal education. Only about half learned to code at school at all. That is not a criticism. Some of the most talented people in this industry are self-taught. The low barrier to entry is what makes the web industry innovative and accessible.
But it also means there is no filter between someone who has spent fifteen years building production systems and someone who finished an online course last month. Both can call themselves a web agency. Both will sit across from your marketing team and sound confident.
Confidence is not competence
Unqualified people entering the industry is not the problem. The problem is that buyers have no reliable way to tell the difference.
The 6sense B2B Buyer Experience Report 2024 found that 85% of B2B buyers had prior experience with the vendor they eventually purchased from. Buying groups place four out of five vendors on their shortlist from day one, and buy from that list 95% of the time. Direct recommendations from peers were the most influential factor.
In plain language: companies pick who they already know, or who someone they trust recommends. Technical competence is almost impossible to evaluate from the outside, so trust fills the gap.
That works fine when the trust is earned. It becomes expensive when it is not.
A marketing team does not know whether their agency recommended the right CMS, the right hosting, or the right architecture. They cannot evaluate whether a server-based setup was necessary or whether a CDN would have done the same job for a fraction of the cost. They trust the recommendation because they trust the person making it.
What dangerous half-knowledge looks like in practice
The server discussion I started with is a small example. But the pattern shows up everywhere.
A company pays € 200 per month for managed WordPress hosting when their site could run on a CDN for € 5. The agency recommended the server because that is what they know, not because the project required it.
A 15-page corporate site gets built on an enterprise CMS with custom server infrastructure, when a headless setup with static generation would be faster, cheaper, and easier to maintain. The agency recommended what they were familiar with, not what fit the client’s situation. We wrote about how to approach this decision properly in our 5-step plan for choosing the right CMS. And when the wrong CMS gets locked in, the consequences compound over time.
Custom backends, database servers, and deployment pipelines for sites that serve the same content to every visitor. Technical complexity that exists not because the project demands it, but because the agency defaults to it.
Estimates that include 40 hours of “custom development” for things that could be handled by an existing tool or service. The Standish Group’s CHAOS research found that only about 31% of software projects are delivered successfully. Half are “challenged” and 19% fail outright. Overscoping is a major contributor.
None of these decisions are malicious. The agency recommends what they know. They genuinely believe it is the right approach. That is what makes half-knowledge dangerous. It does not present itself as uncertainty. It presents itself as confidence.
The cost lands on the buyer
When an electrician installs faulty wiring, there is a licensing body, insurance, and legal accountability. When a web agency recommends an architecture that costs three times what it should and delivers no measurable benefit, the client absorbs the cost.
There is no industry body that audits whether the CMS recommendation was appropriate, no standard that defines what “overengineered” means for a marketing site, and no recourse beyond choosing a different agency next time and hoping the next one is better.
The 6sense research also found that 78% of B2B technology purchases in 2024 were renewals, replacements, or enhancements of existing solutions. That means most buyers are not choosing for the first time. They are choosing again because the first choice did not work well enough. The cycle repeats.
For companies evaluating web development proposals, our guide on thorough evaluation of digital offers covers what to look for in contracts, pricing models, and technical decisions.
The beautiful contradiction
The web industry’s open door is genuinely one of its strengths. No gatekeeping means curious people get in, experiment, learn by doing, and sometimes build things that professionals with degrees never would have attempted. The low barrier to entry is directly responsible for the speed of innovation in this space.
But the same openness means the burden of evaluation falls entirely on the buyer. And most buyers are not equipped for it. They are marketing teams with growth targets and campaign deadlines. Evaluating whether a proposed server architecture is appropriate for their use case is not their job. It should not have to be.
What buyers can do
You cannot change the industry structure. But you can change how you evaluate.
Ask for the reasoning, not just the recommendation. A competent provider can explain why they chose a specific approach, what alternatives they considered, and why those alternatives were worse for your situation. If the answer is “that is how we always do it,” that is not expertise. That is habit.
Be skeptical of complexity. If a proposal for a marketing site includes servers, databases, deployment pipelines, and a team of three developers, ask what would happen if you went simpler. The answer tells you whether the complexity is justified or just default.
Separate building from advising. The agency that builds your site has an inherent incentive to build more. Consider getting an independent assessment before committing to a large project. That is one of the reasons we offer the Headless Audit as a standalone service. We wrote about what a website audit actually reveals and why the findings often surprise the people who commissioned the original build.
Check what they delivered, not what they promised. Portfolios show the visible output. Ask about performance metrics, content management workflows, and what happened six months after launch. A beautiful site that nobody can update without a developer is not a success. The signs that a CMS is holding your marketing team back usually show up months after the agency has moved on. We also wrote about what the first 30 days after a migration reveal and why post-launch accountability matters.
The web industry is not going to introduce licensing exams. And it probably should not. But the gap between trust and competence is real, and the cost of that gap lands on the buyer every time. The least you can do is make that gap harder to hide.