How to start a start-up

Summary

Starting a dotcom company without capital is possible. This article is based on the ideas of Paul Graham in "How to start a Start-up" and on my experiences in Loquo and makes a number of suggestions to potential entrepreneurs.

Introduction

My title is taken from the brilliant essay by Paul Graham, "How to start a Start-up" and in this article I plan to compare the progress of Loquo, from its birth about three years ago up to its recent sale to eBay, with Paul Graham's recommendations (who sold his Start-up to Yahoo). I also add a few ideas (from Alzado and my own experiences) to encourage those who want their own dotcom.

There are two fundamental differences between this article and Paul Graham's.

The first is that Paul Graham focused his article on a generic start-up (in his case a software company), but I am describing a dotcom specifically providing Internet services.

The second is that the original article put a lot of emphasis on start-up capital, specifically venture capital. In contrast, in this article I describe a start-up without initial capital and with the work of its founders as the only investment. This was the case for Craigslist and Loquo and I honestly don't think that these are exceptions.

I would like to repeat that the majority of the ideas in this article are taken from Paul Graham's great essay and I encourage you to read it.

It is hard to start a project from zero, if it were otherwise everyone would be doing it. But at the same time it is not only within the reach of a small minority. As Graham says, there could be twice as many start-ups as there are now and that would be good for everyone.

The beginnings of Loquo

Loquo began its journey as a perfect example of a garage start-up. In its first year it was running on a server in Ubaldo Huerta's flat in the Raval district of Barcelona with a ADSL connection. As a matter of fact, from the beginnings until its sale to eBay, Ubaldo worked from his home.

My contact with Loquo began after discovering the site in a post from Ubaldo in the usability mailing list of Cadius in which he was asking for opinions about the site a short time before opening it. I thought it was a good idea but I thought starting from no advertisements was going to be a problem.

Six months later I saw that the site was beginning to take off and decided to interview Ubaldo for Alzado. We soon became friends and I began to follow the growth of the site more closely, which resulted in writing another article in Alzado. I now knew much more about the guts of the site and was able to write a more interesting article.

Gradually I began to collaborate with the site and from the middle of 2004, after leaving my job in la Caixa as usability consultant and setting myself up as an independent consultant in the previous year, I became a full partner of Loquo.

The site grew exponentially both in numbers of adverts and pages served. Then more collaborators joined, amongst which were some on my friends, until finally a month ago came the sale to eBay, when we all left the project with joy and sadness in equal measure.

Beginning is hard: motivation and demotivation

Every process of starting something from nothing is like being faced with a huge mountain that you have to climb and the only resource that you can draw on is your motivation. The major risk is becoming discouraged and giving up. In such a situation the key is maintaining a high level of motivation and the best way of preventing loss of motivation is to travel light, to avoid being saddled with any heavy expenses that in the more difficult moments could give rise to thoughts of giving up.

Avoiding expenses means calculating what is necessary and unavoidable in order to complete each stage successfully. In this way the goals seem reachable in a reasonable time frame, they don't appear impossible and you don't wear yourself out. If you don't need an expensive Gore-Tex jacket in the first stage of climbing the mountain, save your money and who knows if you will need that jacket after all.

Another of the major risks, which may seem trivial, is planning too much. You mustn't confuse thinking about things with solving every problem at the start. The tendency to meet and discuss everything before doing something can block action, which is why it is often called "paralysis by analysis".

A good ideas isn't worth anything

Paul Graham is absolutely right when he says that a good idea is worth nothing because the difficulty is not in having the idea but in making it work.

The idea behind Loquo is not especially brilliant at first glance, it is a clone of Craigslist, the classified ads system in the United States that was so successful.

Putting it like this is a misrepresentation. First, it is not that easy to be the first to identify a successful idea that has not yet been brought to your market. Second, calling Loquo a clone of Craigslist is like calling Google a clone of Altavista because they copied the text box and the search button.

The idea behind the majority of businesses is simple and it is precisely this that makes them work. From the outside one advertisement site may not be that different from another; the difficulty is making it work. In all sites the principal differences are on the inside.

Another thing that highlights the small value of an initial idea is that a number of clones of Loquo (more than 5) soon appeared, some of which even copied Loquo's own adverts. I repeat, however important the initial idea may be, if you don't make it work well, then whether they copy it or not it is not that important. Loquo continued to do better than the competition, the clones did not affect our growth and Loquo even started to apply the Creative Commons licence to its advertisements so that anyone who could copy them. It was obvious from the start that fighting for copyright was going to be a losing battle.

I don't want to say that the idea is not important. The initial idea is essential and any old idea does not count; it has to be a good idea. What I mean is that at this moment I have something like 10 interesting ideas for Alzado but this doesn't have any merit because the most difficult thing is to get them working. If I wanted to execute one of these ideas, first I would have to get the agreement of the three founders (Luis, César and myself) that the idea was worth while. Then we would have to create the prototypes assuming that Luis could find time in the morning to program, then we would have to refine the idea and correct errors which would be as much in the detail of the idea as in the programming, finally to discover after X months if the idea works or not (and now we are not starting from zero but from 220,000 pages served per day in May 2005). I scoff at the value of my 10 magnificent ideas.

In order to make an idea work the important thing is flexibility. Paul Graham noted that although your idea may be good it does have mistakes and it is critical to discover these as soon as possible. It is no good insisting that you try to make the initial idea work without exploring the unexpected opportunities when they appear. It is no good being afraid of sharing the idea; this is how it can be improved. Forget, "Poor me, they have stolen my idea", it could easily be that they came up with the same idea independently.

The hole in the market and the competition

The other requirement which Graham thinks important is that there is a hole in the market. I'm not talking about a new idea in a virgin market (everyone's dream) but about something that isn't yet being done sufficiently well or something that can be significantly improved.

When Loquo started, the market for classified ads in Spain was dominated by the giants Segundomano and Anuntis (now merged within the French group Trader) and they were not betting on the Internet. Their processes and business models would not work well on the Internet and I suppose that they were afraid of a communications medium where a small team of people could do the same and better without having the dragnet of managing such a large company. Their webs used peculiar processes, for example the user had to send a premium SMS message (1 euro) in order to obtain the contact telephone number of the advertiser. Apart from going against natural logic, that it is the advertiser who pays, the user had to take the risk that the flat being advertised, for example, was still available.

The other aspect that the competition were not conscious of was their placement in search engines. Their sites ranked very poorly in Google which meant that although they had a flow of advertisements from the physical world, they were restricted to the people who already knew the paper magazine and remembered the URL, the image and trade mark. This made it difficult for them to attract new clients and they lost their market share with the growth in the use of Google as the first site from which to search for anything on the Internet. Loquo in contrast was listed in the first results obtained from Google in almost all the searches in which it was relevant.

The usability of Loquo is also good, much better than the competition. Still, I have to emphasize that although usability is essential, to make a web site work it is not sufficient; it is necessary but not the only requirement.

Being listed in the fist results from a Google search we had as many advertisers as we had readers of advertisements without investing anything in publicity. Because we were easy to use, we didn't lose many of these visitors.

I like to emphasize that in my opinion you shouldn't be too worried about the typical statement, "this market is already too saturated; there is no gap". Obviously it is easier to attack a market where there is less competition than more, but it is much more difficult and risky to be the revolutionary discoverer of something completely new than to improve something that you know already works, especially when web design is still in its nappies and much still needs to be improved.

The people and the team

If the idea is not worth much and the value is in making it work, capable people are what makes the difference. It is the people who are the decisive element and not the idea. You can change an idea as you go along, but it is difficult to change a person.

I don't have any hesitation about saying that the Loquo team were exceptional. The internal mailing list for sharing ideas was of the highest standard. But the most important factor was the open attitude and the love of doing things and doing them well.

The formation of the team at Loquo was unique owing to the characteristics of the project. Loquo didn't have any income, neither did it have a registered company behind it; it was no more than a group of people brought together by the founders Ubaldo and Lissette and later by the collaborators.

There was no process for introducing in new people; every case was different, but everyone agreed what a good idea Loquo was and had the same desire to learn how a project like this could operate on the inside. In the majority of cases, the collaborators looked for Loquo and not the other way round. Being spontaneously interested in Loquo was the best endorsement for joining the team.

In general, each collaborator opened up a new town, the town where they lived. There was no obligation, it was a collaborative project "in the style of Linux", everyone worked as much as they wanted and when they wanted. The large number of people involved in Loquo was something specific to the Loquo project and I don't believe its decentralized growth by towns is something that can be extrapolated to the general project.

Graham recommends that the ideal number of founders for a start-up is between 2 and 4. There is too much work for one, but too many people complicate  decision making. I would agree that 2 or 3 are the perfect number.

Although the collaborators of Loquo may have seemed gullible in joining a project without a formal contract, the common denominator of everyone was the desire to learn. We were not losing anything nor were we obliged to do anything; it is natural that some did more and others less. It was also natural that each collaborator's available time was variable. It is inevitable that who has the most free time works more than the others; its like swings and roundabouts.

Although it may seem strange, I think that this arrangement can be extrapolated to other projects. In my personal opinion, to start talking about distributing profits when there is no income and of obligations when there are no benefits does not make much sense. If you don't trust your collaborators, perhaps you should ask yourself why you are in a project with people you don't trust. Clearly, if someone puts in money at the beginning, everything changes (there are contracts and all the rest), but in a project starting from zero and without resources, the flexibility based on mutual trust is vital.

In my case, as independent usability consultant, I was able to implement many of my radical ideas on usability that the majority of my clients would not have dared implement, putting them to the test in Loquo and its millions of monthly page visits. This experience is literally worth its weight in gold for my next project and my job references.

Personal profiles and the birth of the team

Paul Graham talks about the need to have technical experts amongst the founders and I am in complete agreement. If you don't have a programmer in the team of founders, then you can forget it. Thinking about technology as if it's something that you can sub-contract or delegate to someone who is not at the heart of the team does not work. This is simply because if you don't understand the details of the technological consequences of every decision, you will have a problem no matter how much they explain it to you. And this is not counting on the agility for making changes that is possible when one of the founders is a programmer.

I don't mean that a project must be controlled exclusively by the programmers; normally programmers have a hard time understanding what the ultimate client really wants. What I mean is that it is often easier and faster for a programmer to learn to understand the needs of the client than for an economist to learn advanced technological concepts.

In the case of Loquo, the two founder members, Lissette and Ubaldo, were programmers and the majority of the collaborators were very strong in some technical area.

The business model

In my opinion, the business model is not the most complicated thing about a dotcom and should not be the thing that concerns you the most at the beginning.

First concentrate on creating something good and making it work, then be patient and allow things to take their course. If the web site works, if it has a lot of traffic, the business model will make itself evident.

I don't mean that you should completely forget about the business model as they did in the time of the bubble. You have to have an outline plan, but to have a detailed business model from the beginning (unless you have an online store) is very dangerous because it is most likely that it won't work and your predictions will be mistaken.

Until you observe your users using the site for a while (at least a year, but I would say two or three) it is very difficult to know what services they will be prepared to pay for. It is more usual that with the passage of time the business model will emerge little by little in an almost spontaneous way.

In Loquo, although there was not enough time to implement it, the business model became clearer all the time; you only have to take a look at the site and the forums to know what it is. There came the time when some users demanded to pay in some sections, which pleased us a lot. Others muttered that Ubaldo was a communist because all the advertisements were free in Loquo. Others foresaw that our intention was to make people pay in the future.

Although I have no idea what will happen in the future with Loquo (now that it is controlled by eBay), I a convinced that on the Internet everything or almost everything must be free to be successful. The success of El Mundo is due precisely to this realization and they balance their books by income from publicity as its director, Gumersindo Lafuente, told us and they will never charge. The big mistake of El Pais was to charge for everything, a decision that they had to reverse two years later. You can charge for specific premium services but never for the main content of the site.

Along these lines, approximately 95% of the advertisements on Craigslist are free; they only charge for job opportunities. In addition, "Jobs" is not any old category; private individuals don't make job offers, so it affects only companies.

In my article, "Don't design your web to make money" (No diseñes tu web para ganar dinero), I explain this in principle detail. The main service of a web site, its core and 95% of its contents must be free. The business model, the services that are paid for, must be a small part and should appear later when they are requested.

The people who think that sites that are free or almost free will charge for everything when they reach a de facto monopoly are mistaken. The great advantage of the Internet is precisely that when the competition is one click away it is difficult to create monopolies because there is almost no cost to making a switch. How long would people wait to move to another search engine if Google stopped working well? No one would have to install or un-install anything. In a few months Google would be sunk; recently Steve Ballmer, CEO of Microsoft, said that, "Google might disappear in the next five years!". On the other hand, Microsoft has a product that is physically installed on computers with all the ties that this involves, so that if this company went down tomorrow its operating system would continue to be used for years. I am not saying that any of this will happen, but from the perspective of Ballmer, Google is very much at risk.

Time and Money

Although Paul Graham talks in his article about how to get venture capital, I am not going to talk about getting financial resources to start a project. Firstly because I don't know how to do it, secondly because I have seen some cases where a company has started without initial capital (Craigslist and Loquo) and thirdly because I think that in Spain it is not as easy to get venture capital as in the US. I don't know about you but venture capital for me is a far off dream, like science fiction.

Not having any money is an important drawback, but not insuperable. Graham himself says that when you obtain investment capital, what you really get is time in the form of salaries to pay for hours of work. In other words, even if you don't have money, if you have time, you already have capital.

But be sure to find out if you really have that free time.

I also assume two things:

a) You either hate or don't know anything about the bureaucratic paperwork that goes with getting subsidies and public grants. In my opinion the worst is not how complicated the whole subject is; it is simply a discouraging use of time writing boring reports for a project that you don't know is going to be accepted.

b) You don't know how venture capital companies work nor do you have any contacts with them. As I said before, I believe that in Spain there are few such firms and it must be difficult if you don't have any contacts in this world. In any case if your business prospers, I am sure that one of them will approach you with an "offer you can't refuse".

Therefore, if your only capital is your time and the time of your co-founders, you have two options:

a) Put on the table the whole of your capital, dedicating your whole working day to your project. You will have to live off your savings for between two and three years, which is risky, but less risky than having to ask for a loan that you don't know you can pay back. In the worst case you can always rejoin the job market with very valuable experience.

b) Dedicate your free time outside of your working hours. The project will progress more slowly, it will be limited in scope and your relationship with your partner/family may suffer. The advantage is that you don't put much at risk.

An intermediate option may be to start by investing your free time and then if the business takes off, to dedicate your full time. The concept of "free time" is a bit ambiguous. There are people who are capable of working more in their free time than in their working hours and others whose working time has gaps. In any case you are not going to go very far if you don't put in a lot of hours.

With respect to the money that you need for the physical side, you already have the two essential resources, computer and connection to the Internet. To begin, the only other thing you need is space on a server. It is not at all crazy to start with a hosting plan for 9 dollars per month with Hostgo or a similar company.

The important aspect about spending little is not the money exactly. It is important that when the hard times come, when you see no progress in your work, you don't feel that not only are you not getting anywhere but also that you are losing money. To be able to say to yourself, "OK, I have not made much progress but I also haven't lost any money" helps to maintain your motivation until you are through the bad patch.

In Loquo, cost minimization was an art. At first, the Loquo server was in Ubaldo's home connected to an ADSL line. To avoid filling up the disk and slowing the response time only one picture per advertisement was allowed and only in some sections. Ubaldo worked a full day in his home from the beginning up until the sale to eBay, putting all his capital (time) on the table.

I recommend that you avoid any fixed expenditure at all costs. Having a dedicated server at 150 euros per month may seem to be an acceptable expense at the start, but all initial fixed costs give rise to problems. As soon as you incur another fixed cost of 150 or 200 euros per month you will start to feel the pain, and although you may still be able to pay, it will lower your motivation and increase the feeling of, "I am throwing money into something that I don't know will work".

To the people thinking of starting an on-line shop, I would say that you should avoid the fixed costs of warehousing and stock management. Why not use the stock in a physical shop that is already selling? In this way the on-line shop with your trade mark is another client for the physical shop that takes care of the boring details, receives orders and sends them out. Until the time when you develop a large sales volume your prices may not be the cheapest, but the only thing you lose is the work of setting up the web site. You will have all the time in the world to make changes and improvements without losing money every month. Obviously you will have to work hard so that your store is better than the others, it won't be easy, but there are many areas where the on-line stores are really poor.

Grow slowly

Paul Graham says that it is the circumstances that make you grow slowly and this is a good thing. This also happened spontaneously with Loquo.

Collaborators joined us little by little which gave them time to be absorbed into the culture and understand the way the web site worked before new ones joined the queue. Because no one was paid for doing the boring routine work, we all had to do everything, including moderation and technical support. Moderating a site like Barcelona with 2000 new advertisements a day gives you more information about how people are using the site than any explanation would give. Answering technical support enquiries helps you really understand what problems people have with the site. Those who have the money to pay for technical support staff don't know what they are missing.

Graham talks about the good image a project gets when the programmer personally answers support calls and can solve the problem, or promise a solution without prevarication.

In the case of Loquo, people didn't call often nor did they send many e-mails (we think that this was because of the good usability profile of Loquo), but Ubaldo personally answered the telephone. You could also see the photo of the local collaborator who answered in a few hours the e-mail that you had sent.

In a project of this sort the telephone and e-mail are the basic elements that generate confidence; they are your method of contact with the users. The photos of the team are important. Not having an important brand is also an advantage. The popular brands try to show the human face to their businesses, but you don't need to do this as you are already an individual; you only have to let your users/clients see you as you really are.

Things to bear in mind

Detecting mistakes

Graham says that you have to remember that you are going to make mistakes and that sometimes you are going to put your foot in it. For this reason, the key is to notice and correct the errors as quickly as possible. The solution is not to discuss (digress) a priori for months. However much you discuss your errors you are going to make similar ones later. It is better to create an initial prototype as soon as possible. Its easier to detect errors in something physical in front of you than in ideas floating in the imagination. It should be possible to build an initial prototype in no more than a week, perhaps even a day; if you need more time, perhaps it is not a prototype.

The weak link

If your model has a physical presence, a contact with the real world, be careful because in this case you are not solely responsible and you are at the mercy of certain people behaving as you hope they behave.

For example, if you base your communication on e-mail, remember that in Spain many people don't respond to e-mail immediately and others don't read it every day. If immediacy is important for you, perhaps e-mail is not the best communication method.

It is also dangerous to assume that the professionals in a given sector will radically change their method of working. People have their own way of working and it is easier for them to make a small change than to switch over to a different method (the one you propose), however good it may be.

Big Clients

Trying to win big clients is attractive, but complicated. Often they will be won over by some one with more commercial clout than you. Of course to gain access to these large companies seems ideal, all your problems solved in one go with a contract figure of many zeros, but in practice it's a lottery. Paul Graham mentions that he prefers small companies that are always more accessible.

Understanding growth

Don't overestimate the traffic to your site in the short term, nor underestimate it in the long term. If the number of visits to the site doubles every three months, even though two times a small number remains a small number, don't be discouraged because you are progressing well. In the long term a growth of this sort will rapidly exceed the predictions because in the short term it seems as if you are not making any progress at all. If you start a new site from zero and your traffic does not double (as a minimum) every 6 months, then for sure you have a serious problem.

Cost of sales

There are business models that make sense but require large investments/much time so that the benefits may be marginal if later the turnover is not high. For this, the Google's AdSense program is so useful. You can set up a Google Ad in 10 minutes and there are no management or maintenance costs; all benefits.

Excessive initial complexity

Graham recommends that you are not too ambitious and start with projects that take on too much. Improving something big is complicated, but improving something small is relatively easy. The niches in the market place are often the simple practical things that could work better. Also, if you make something small, it is simpler to use, cheaper and so it's easier to convince people of its value. Later you will always have time to add functionality and features when this becomes necessary.

Generating content ad infinitum

Undertaking to constantly generate content for a web site with a small team is a model fraught with problems. Any site needs as much content as possible to generate traffic from Google, but doing this with a small team is too tiring. The simplest option is to create a community, to allow the users to generate content, to post articles or news, to publish their opinions about products, to create discussion forums etc.

Opportunity cost and return on investment

"Opportunity cost" is a concept that a lot of people don't understand in Spain, and the term is a bit confusing. It simply means, "wouldn't it be better to use my resources on some other project"? When you do one thing, you leave others undone; the question is, would the things you aren't doing be more or less profitable than those that you are doing.

Opportunity cost is especially important when you have a whole bunch of things to do and you have more and more ideas coming. It's what often happens in these types of projects. When your only capital is your time, not managing it well is the same as losing money.

Keeping in mind the return on investment of each task is the key to concentrating on the most important task of the many you have pending.

Legal considerations

Like on the point about accessing venture capital, I have no idea about the legal considerations that apply to web sites (nor do they interest me). I expect most of the readers of this article are in the same boat. My opinion is that at the start of a project you will spend as little time thinking about them as you can get away with.

If you don't have a business model at the beginning and there is no money involved, it is not necessary to start up a company or set up as a sole trader. As regards complying with the Ley de Servicios de la Sociedad de la Información y de Comercio Electrónico (LSSI) it is sufficient to supply your personal identification number (NIF), a physical address and some kind of privacy undertaking.

Other examples of projects

As the purpose of this article is to encourage you to start up a project, to show you that this is not something unreal and far off, I am going to give some examples of people who have already done this and succeeded.

Being a consultant and taking advantage of the dead time between developments, taking advantage of code that you already have, to start up something new makes a lot of sense.

This is what they did at  Evolucy, creating  Caldoso in a couple of months. This is a search engine designed specifically for the  Boletín Oficial del Estado (BOE) which simplifies the finding of information in this official document set. It should simplify the lives of lawyers, consultants and companies and we will soon know if it really has.

Buscavuelos finds the cheapest airplane ticket for you, including the low cost carriers. It was developed by  Oscar Frias who has developed other sites as well. I have tried it and it works well; soon we'll see if other people can be encouraged to use it.

David de Prado of Dnxgroup has created an online store, iAlbea , which sells luxury cases for the iPod. It is in English and targeted at a very specific market.

Toni Vicens, who works in the IT group of the  Mutua General de Cataluña, has launched into the theoretically saturated online market of selling wine with Vinissimus. It has started well and I am involved with it and soon I will tell you has its going.

If you know of any other start-ups and dotcoms of this type, please send me a link and a short note and I will add it to this article.

For more translated articles, visit: Translations

Return to Extracts from Planeta Código