A 500,000-Euro Website

Can a website be worth 500,000 euros?

Of course it can. But then one day you wake up and read that the new Spanish Senate website cost the stunning amount of €448,819.25. Other outlets put the number at around €437,691.50.

If you are not a professional in this industry, that number probably gives you chills. If you are a professional, the first thing you probably feel is not horror. It is skepticism. Because when a project costs that much and the result is clearly mediocre, people inevitably start asking the obvious question:

Were we ripped off?

And once that question appears, the next one follows immediately:

Where did the €448,000 go?

Half a million euros for a website? Excuse me? I would have made that same website for much less. Someone clearly pocketed the money. This has to be some kind of laundering operation. What an insult.

That quote pretty much captures the general mood around the Senate website. Anyone who sees it asks the same thing:

Does this really look like a 500,000-euro website?

I do not want to sound like a government spokesperson defending the indefensible, but I feel the need to dismantle a few beliefs that many pseudo-professionals, internet consultants, and loud voices in the Spanish internet scene have about cost, value, and, above all, production.

Capacity

This? I could build this with one hand tied behind my back.

No, my friend. That is not how this works.

Nobody builds the Senate website alone. At least, nobody in their right mind does. You only need to read the project specs to understand that this could never be a one-person job.

The Senate website is a monster. It has documents, categories, archives, processes, and an enormous pile of old institutional material dragged along by years of mediocre consulting work and public-sector disagreements. It has been a mess for a long time, and it will probably remain one for a while.

Based on my own experience working with government projects, if you tried to build this alone, instead of finishing it in six months, and then delaying it another ten, you would probably finish it around the time the tenth-generation iPad comes out, Catalonia has become independent, and Lionel Messi is retiring from soccer. In fact, the project was awarded to three companies. A fourth company provided the underlying technology. That alone should tell you something.

Valuation

We could do this project for €100,000.

Maybe.

But pricing is always a funny subject because everyone thinks their number is the real number. If a professional in this field earns at least €25,000 to €30,000 a year, and the project is planned for six months, then once you include the consulting company’s margin, the size of the website, the amount of material, the sections, the bureaucracy, and all the surrounding work, the total does not sound crazy. In fact, given the scale, the money seems low and badly managed.

There is also the subjective side of every professional project. You cannot simply say, “I would do the same thing for €100,000.” That is your price. Mine could be €2 million. It is like looking at a Pollock painting that sold for €100 million and saying, “I could paint that for €1,000.” Talent has a price. If you have forgotten that, take a walk and clear your head. Building a website is not just sitting down and writing PHP, HTML, and CSS. It is more than that. You have to identify the problems and find the best solutions. Good professionals cost money. And they should. I do not believe €100,000 is enough to produce something truly great when we are talking about a large-scale, high-responsibility public project.

According to the Senate’s own numbers, roughly half of the total cost went to licenses for the content management system, Oracle UCM, electronic processing infrastructure from Kinamik, search licenses from Oracle SES, and a new Oracle WebLogic application server.

So a big part of the money went straight into licenses.

And that raises the next question: Why did they not use free and open source software?

Hey, if they had used open source software, it would have been free.

Yes, open source software can be free to acquire. But implementing it is not free. Nothing installs itself by magic. Someone has to configure it, adapt it, integrate it, test it, secure it, deploy it, document it, and maintain it. That means people. And people cost money.

According to the Senate’s IT manager, the average cost of proprietary search licenses was €109,000, while open-license alternatives came in at €111,000. That number honestly leaves me stunned, because there is an inconsistency there. If the software is free and open source, acquiring it should not cost anything. What you pay for is the service around it. With proprietary software, however, you pay both to acquire it and to install it. So either proprietary software vendors were giving away the product and charging only for installation, or the person responsible for evaluating this got something wrong.

Looking more closely at the issue, I have been told, and someone can correct me if I am wrong, that the specs already required Oracle and Java with no exceptions. If that is true, then the statements made to the press about evaluating open source sound like pure smoke. Open source was ruled out from the beginning.

That is a shame.

On the other hand, this reminds me that the proprietary software industry is much better protected in these cases. These companies lobby constantly, and they usually have stronger certified-developer networks than open source alternatives. There are more certified technicians for some proprietary systems than for comparable free systems. That happens everywhere. I have competed against big companies, and I know what it is like to see them offer work almost for free just to get inside and sell licenses in the next phase.

So, in the end, around €250,000 went into licenses instead of labor. And labor is where the money should have gone.

The real problems

At the design level, there are plenty. For starters, this was not designed. It was made up. And I do not mean that as an insult to makeup artists, who often do excellent work in the arts.

What I mean is that the design is not functional. It does not seem to follow the principles Dieter Rams famously defended: innovation, usefulness, aesthetics, clarity, restraint, honesty, longevity, attention to detail, economy, and as little design as possible. The design clearly feels like it came fourth in the list of priorities. I can guess why: they probably did not really invest in it. Otherwise, either the website would be properly designed, or the people involved were not designers. It is that simple.

This design lacks many essential qualities of design itself. And if I am allowed to be subjective for a moment, this is a design to forget, not to remember. It also shows no evidence of serious UX or UI analysis. I see large, predictable mistakes that make it hard to believe this was properly studied beforehand. Take the breadcrumbs, for example. They are awful. They repeat themselves constantly and ignore basic hierarchy and redundancy principles. They look like they are there just because someone thought they had to be there. Font sizes, information architecture, color systems, general functionality (I could spend the whole afternoon listing problems).

And when it comes to details, I cannot be brief either. There are dozens of defects, and every one of them damages the credibility of the Senate itself.

At the code level, there are also several problems. Some are serious. There were small vulnerabilities that people used to inject text and other nonsense, leaving the website exposed for days. That kind of thing shows a lack of preparation before launch. It reminds me of The Checklist Manifesto: basic checks matter, especially in projects with public visibility.

There are also many amateur-level mistakes.

Javier Ramírez summarized many of them, although not all, and some have already been fixed. His post, “Los problemas del desarrollo web en España resumidos en senado.es”, is worth reading.

What should have been done

They gave a facelift to a website originally created in 1997. Instead of applying makeup, they should have investigated the best way to build this kind of public website for the present day. Adding Twitter to the homepage, or to every senator’s page, is not innovation. Anyone who believes that should stop wasting time and see a doctor immediately. The site needs to be rethought. It should be built intelligently so it can last for decades and survive the countless legislatures that will come after this one.

In my opinion, the best thing to do is invest in research. Not patch together something that should be genuinely redesigned from the ground up. This experience should also be used as a model for other government projects. That is the only way to stop acting foolishly and start spending public money properly.

There should be a serious investigation committee that forces the awarded companies to comply with what they were paid to deliver.

Second, research should be commissioned from several companies, removing the prejudice that only large multinational consulting firms can bid for this kind of work. That mentality only ruins opportunities for smaller companies and for people who are actually innovating. The public sector also needs to move away from proprietary software once and for all. Using proprietary software is like buying something on installments: you never really finish paying for it. You are always stuck in the same cycle. The money should go entirely into development, not licenses.

Imagine a world where you had to pay a license to use HTML, or where a website using CSS required a special paid permit. Costs would always double. Using proprietary software in this sector does not just tie you to bad practices. It also reduces your room to maneuver. Open source software, by contrast, lets you change everything if needed. This is not a new discussion. Rivers of ink have already been spilled on this topic, and that started more than ten years ago.

You also cannot give the entire development process to one company. Responsibilities should be split.

Design should go to a company that specializes in design. Frontend code should go to another. Other areas should go to other specialists. That is how movies are made: work is distributed among many companies that are experts in specific things. The Senate website is only the most recent and painful example of what is still happening today. With all the companies involved, it will stand as a clear example of how to do things badly with public money.

But the most shocking part of this whole story is that, in 2012, we are still repeating bad practices from the previous millennium. And in the middle of a crisis, spending public money on something badly made only makes society angrier. People are tired of seeing no progress.

They are tired of watching their contribution wasted on things that do not really help.