Ricardo Galli by Benjamí Villoslada Gil
Ricardo Galli quite openly tells us the most intimate details of Menéame; how much they earn monthly with Adsense, server costs, purchase offers, internal architecture, keys to success, problems, Varsavsky…
It is not very often that somebody explains so much about their own project. A brief summary and interview follows:
- Menéame runs on a single server, in spite of having 65 million pages viewed according to Webalizer y and 1922 position in Alexa..
- Menéame generates 4,000 dollars monthly in Adsense revenue
- The current expense in servers is 500 euros
- Ricardo Galli prefers Adsense to any other advertising proposal
- He believes that it’s success is due to the fact that the software was developed by them from 0 and published as open source.
- They accepted Varsasky as an investor to cover costs and to feel more secure (sues, legal issues, etc.)
- There are only two developers: Ricardo Galli and Benjamí Villoslada
- Initially Varsasky had a 10% share? of the company and now it has 33%
- They have received offers for the whole of Meneame of more than 6 million euros.
Eduardo Manchón: How much did it cost to start up Menéame? What are the current costs? Have the costs been a problem for Meneame?
Ricardo Galli: The only initial expenses were the domain names and the rent of a virtual server (on Xen). The first server cost us 40 dollars per month. We then had to extend it to about 80 dollars, but we had many problems with the Xen, it hung with high loads. So we had to migrate to our first dedicated server in ThePlanet. It cost us about 150 dollars. Now we have used Ferca for almost a year.
After that we did not spend anything else, except for the expenses of notaries, lawyers and consultants when we began with the SL(1). We created the SL for two reasons. First due to legal issues, mainly related to the LOPD(2) and the LSSI(3). The second reason is that it is “the smallest” reasonable option to accept external investors. As an anecdote, the contribution of capital to the SL were the domain names, we did not contribute in cash.
For both servers in Ferca we pay little more than 500 euros IVA(4) inclusive. I believe the maximum of monthly traffic per each server is about 2,5 Terabytes, but I’m sure we are very far from those limits because almost all the content is text and the server compresses the text before sending it – if I am not mistaken, when you accede for the first time to the main page you download only about 30 KB altogether. I believe that it is almost a Record 2.0; -)
Eduardo Manchón: In Menéame you have Adsense, I believe that it is your only business model. Does it cover your expenses? Does is it give//bring? you benefits?
Ricardo Galli: At the moment we have an income of a bit more than 4,000 dollars a month from AdSense –and it is increasing every month–, and a few more from campaigns in FeedBurner and others that are sometimes contracted directly.
It covers expenses because we control ourselves a lot, we only spend if we can compensate it.
We had many offers and we even tested other advertising systems, but we always found that it is time consuming and requires using advertisements that are more intrusive than what we considered reasonable. But we are in contact with the publicist to see if they can take care of it as long as they assure us to surpass the income from AdSense by enough of a margin and the advertisements are not intrusive.
We are convinced that advertising will provide far better revenue than currently, AdSense and all advertisements in general. In the US it is said that the online portion of the advertising pie is still small in comparison with other media, imagine that we (Spain) are significantly below that. In fact if we had the same proportion as sites like Digg we could have revenues of about a 4-5 million euros a year. Now I remember that blogger that had shown the Achilles heel of AdSense during a month and made almost 200,000 dollars. A few days ago he published a screen capture of the statistics, and funnily enough the impressions and “click ratio” were very similar to Menéame. Nevertheless the eCPM and totals were bigger, 25 times more than ours. That can only give us hope : -) [*]
[*] Of course, we realize the thematic differences between the sites, his is dedicated to ringtones and ours is too general. Even so, the difference is terrible.
So now we prefer to dedicate time to other things and not to bother the users. The time will come when we improve considerably, at the moment, we do not have the urgency. Our “burning rate” is almost zero.
Eduardo Manchón: What is the infrastructure of Menéame at the moment?
Ricardo Galli: Now we serve everything from only one server (a Dual Duo2). We have another server that we used to replicate the database, backups and those processes “off line” that require lots of CPU and disc usage. It is also prepared to balance the load with the first server or to replace it immediately, but we have never needed it.
Then, of course, we have what we used in the house and the “total connectivity” that Benjamí and I have. We go everywhere with laptops and connection 3G (Vodafone). It is essential for us.
Eduardo Manchón: What have been the main problems that you have had in the life of the project? How have you solved them?
Ricardo Galli: We had many problems with Xen and we could not find what was causing them. The AdSense income was not quite enough to pay a dedicated server and we did not know either if Menéame was going to grow more or to die in the short time. It seems a triviality, but it was hard “to risk” little more than 100 euros a month to start “for real”. That’s why I say that we are not “enterpreneurs”. Looking back, it was silly to doubt so much… not to have started with a dedicated server from day 0.
At a personal level I had a bad time when we were accused of censuring or manipulating votes. Although I dedicated a lot of time to clarifying things, it seemed to be in vain. I am used to it now, but I felt stressed and an incredible lonelyness, as if the world was falling apart and nobody was helping us (although it was not true). With hindsight, it seems silly and it is just something that I suppose happens to all of those who begin projects that then become pretty popular.
I also was quite stressed due to security bugs in the software. I am not an expert in this matter and had never made a program used by so many people, or one so popular. It is stressful. Luckily, since it is free, I received many “patches” for the security problems, in particular from Alex Concha. These patches helped us avoid more serious problems. Now I can say almost with pride that for many months it has been “rock solid” (probably somebody after reading this will warn us of some serious bug, it always happens : -)
Eduardo Manchón: Could you detail the scalability problems that you have had?
Ricardo Galli:In regards to the escability issue, systems such as Menéame or Digg are quite special. Their bottle neck is mainly the queries to the database, and then the compilation and execution of the code. If the database is badly designed or the system does not have appropiate indexes, the system will not work because it will be overloaded and will hang almost as soon as a “heavy” query is executed.
At the begining Menéame did not have all the queries properly optimized. In our computers at home they worked very well –the data base was still relatively small– but as soon as they were in Xen, and with many concurrent queries, they generated a lot of input/ouput to the disc and created many Apache processes. That caused our server in Xen to have many problems (too much time of kernel threads blocked by I/O), even hanging our guest core. Soon we saw that there were more people having problems in very high load situations –they say that now it is solved–, so we decided to change to a dedicated server. Then all problems of this type stopped.
In regards to scalability in general, my “hobby” as a programmer is to write more efficient code — for example WP-Cache –. I spend lots of time analyzing the code and the queries. Perhaps if I had not been so careful, we would have needed more than five servers to support the present load.
Demanding such efficiency from the software limits to a great extent the functionalities that can be added, and also means that we can not use abstraction layers –as templates–. On the other hand it eliminates the related problems of scalability and their costs and problems (as well as entertaining me : -)).
There are people who think that the programming time is much more expensive than adding servers. However not everything is so linear, scaling by adding servers increases the complexity of the system, causing the administration to be more complex and fundamentally increasing the chances of failures or “latencies” that are more difficult to solve.
I am fan of KISS (Keep It Simple Stupid), and I believe that time should be invested in having a very efficient and simple architecture and code for “social networking”. The benefits can be seen in the long run, especially if you are beginning a company like ours and your resources are very limited while at the same time you need to grow desperately.
Eduardo Manchón: At one time there were several projects similar to Menéame, but few of them have prevailed. In France or Germany there are no equivalents to Menéame with the same level of success. What do you think has been the main reason for Menéame’s success? What have you done that similar projects have not?
Ricardo Galli: Puf… difficult, I do not know if I know the answer. I will say one thing.
I believe that the fact that the software was developed by us from scratch and published as free is perhaps the most important factor. It gave us what it is known as “principle of authority”. It was also important in order to resist the usual accusations or conspiracy theories put forward by the public. It is like saying: “here is the code that does everything, sure it has errors, but they are not intentional nor aimed at anybody in particular”. More or less like Google’s PR algorithm, which makes big “fuckups” and sometimes its changes harm a lot of people economically, but it is ok to blame the algorithm, it cannot question you nor answer back to you through a blog : -)
I think the fact that we did not take ourselves too seriously or be too “up ourselves” in front of the public had a lot of influence–although we did take it very in seriously underneath, as if our lives depended on it–, the provoking name, the mascot that looks like a stuffed toy, the white page style with big fonts. We were trying to be honest and respect the vote of others (although we didn’t always agree), sometimes publically acknowledging big errors… I suppose that all that gives an important personality to the project.
Also I believe that it helped a bit that Benjamí and I have been bloggers for long time and have participated in several communities and mailing lists. Still we make mistakes, but it helped us to avoid making again the millions of serious mistakes that are made when you interact with many people and you do not have enough experience –for example, being too serious about yourself or your project.
Eduardo Manchón: Principle of authority… Isn’t this too abstract? Isn’t it that creating the technology from scratch gave you a dominion and absolute knowledge of the technology used that is essential to succeed?
Ricardo Galli: What you say is part of the principle of authority. If you have developed all software from scratch, it actually works, many people use it, and also you release for free, it at least means that:
1. you have a good command of the technology,
2. you know how to adapt it to the “real world”,
3. it is adapted for its users,
4. and in addition you do not have problems or are ashamed of showing and sharing the code.
What I just said are usual things in open source software, although newbies do not value its importance and benefits. I hope that if somebody takes the Menéame project as an example, they also consider the importance of the open source existence that has allowed the project and to work with the same philosophy.
Eduardo Manchón: In spite of already reaching such heights of success, I do not believe that you are going to stop. Which is the main challenge that Meneame faces right now?
Ricardo Galli: Right now it is “to professionalize”. Fundamentally the advertising and from that to mount a team that helps us and leaves time for us for other things, plus the employment of developers if they are required.
On functionalities… we have passed the phase in which everybody requested many changes or new important functionalities -and we have paid lots of attention—. Now it is quieter. Although I do program every day, there are details or programs that are non-visible, like those that automatically detect abuse.
We have some ideas that we will be implementing little by little, some soon, others in the long term: messages and private messages; geographic queries; a bot that detects all related news and news that speaks of the same subject as news already published; personalization of individual categories; a more formal API and with more functionalities; warnings messaging through SMS; more interaction with mobiles, etc.
In general we are obsessed that everything we implement revolves around the “news” and connects and sends visits to external sites. We do not forget that that is the core of Menéame.
Eduardo Manchón: What is your role and what is the role of Benjamí Villoslada in the daily routine of the project? Are there more people involved?
Ricardo Galli: Benjamí in addition to being in charge of the graphical design of Menéame, is a kind of “CEO” and finance director. I’m too disorganized. He is in charge of the administration, the accounting, taxes and seeing that all the bureaucratic activities are up-to-date. He also speaks with publicists and even with TV producers that are interested in making shows about Menéame.
I am more focused on the architecture, development, systems management and being the person to give and receive blows, but also to receive all the merits: -)
We both participate a lot in “the fisgona”, comments and we continuously make sure that everything works and nothing breaks down–it takes a lot of time–.
There are 20 volunteers who help us a lot with “admin” tasks, in the sense that they are there to help people with their submissions, to ensure that users are not insulted by comments, to detect spammers, etc.
This week we will finalize the contract for a part-time person who will help us with that and with access to the server –before we could not do it due to the LOPD– where there is a series of applications to control abuse, like spam with clone users, abuse of votes with different accounts and restarting routers, etc. We have predicted that we will incorporate more than two people in the same role very shortly. The problem is that since it’s the first time we’ve written a contract the procedures have been time consuming.
Eduardo Manchón: Martin Varsavsky is investing in Meneame. Did you look for him or did he look for you? In your posts you also mention other investors, friends and family. Do you have other investors?
Ricardo Galli: He looked for us, he was not the only one, but he has had infinite patience. Not only did he offer excellent money, but also confidence, advice and contacts to solve problems more quickly than other ways which could have taken us much time and effort.
Also, he does not bother us for anything and is always willing to lend a hand. When we speak or we exchange ideas, he knows what he is talking about and the mysteries surrounding the “audience” of Menéame. I would say that we could not have a better partner. Given the discussion level that we had had in the past, I would say that we were doubly lucky. I do not know if we deserve it: -)
Eduardo Manchón: What are your objectives when attracting investors? To cover costs? To get advice or other type of aid?
Ricardo Galli: To cover costs; to have funds; to create an infrastructure; to have protection so that if we are sued –like Frikipedia– we will be able to have access to the best available lawyers; and mainly the personal economic security that we are not wasting the time and losing lots of money (otherwise we could have done other projects that did make more money).
Eduardo Manchón: For somebody with a technical profile that has a project or wants to create one, when would you say is the right moment to open the door to investors?
Ricardo Galli: No idea, in our case I believe that we had pure sheer luck. Even the way we did it with Martin was very good. At the begining we only gave him a 10% share in the company that allowed us to have total independence and to feel that the project was still ours and personal.
Since he had behaved and we wanted to be fair we offered that he could have up to a 33% share–keeping the same price– when he desired but only if he saw that the project evolved accordingly. He made his option effective a few months ago, which gave us more confidence.
I would say that to sell a small part first to have some money can be helpful without you losing the passion for something that is “yours”. You won’t lose much either if you have failed in the estimations. Soon you could sell a bigger share when you have more information and experience.
Eduardo Manchón: Have you had more offers of the whole of Menéame?
Ricardo Galli: Several, a few for more than six million euros.
But we do not make a decision for several reasons. One is that Martin says that it is worth or will be worth more money. The other fundamental reason for us is that we want to stay involved in the project, we hope that it remains known for its independence from corporate interests and also that it entertains us and that “commercial objectives” don’t opress us. We have all that today although we earn peanuts, but it is more difficult to obtain it when somebody has invested millions in the project.
Let’s see if someday we can bring all this together. It does not worry us either.We are not in need of money nor are we aiming to become rich overnight.
We are also convinced that the advertising income can and will increase, as the audience will too (in a year we have tripled the public, we guess that for next year we will have grown between 50 to 100%). If these forecasts are right, even in the “worst case scenario” we will be able to maintain Menéame ourselves and with good profits.
Eduardo Manchón: Companies say that university does not prepare students well enough for the market. In your case, as university professor and administrator of a project in the real world, do you think that they are prepared for the labour market when leaving university? What about for starting their own project?
Ricardo Galli: Yes and no. I’ll explain this according to what I understand although it may sound provoking.
Firstly, with regards to “technical” matters or topics that are taught in university, they are not bad. They can be improved, but they are reasonable. As far as the technical capability of the professors, there is a bit of everything, but I can assure you that most of my colleagues are better professors and researchers than I am. Sometimes “luck” lies not in the programme but in who has been your professor in key papers. Maybe you had somebody brilliant in one paper that inspires you in such a way –for example showing what research is still to be done– that they give you the desire to begin a related project, or that you find its technical or commercial outcome.
The problem lies in the post graduates degrees, in the access to researchers and businessmen with experience. The papers that are lectured in engineering or bachelor degrees do not vary too much with those taught at “California Valley”. The professors and contact with the “scientific community” that the university offers are different. For example Google or Yahoo started out from postgraduate degree students of Stanford, where they have had the opportunity to work with and to receive advice from gurus -Page and Brin were mentored by Hector García Molina, a prestigious scientist of distributed computing–.
What does this offer them? First of all lots of nerve, ego and self-confidence. And that is the main problem that I see in our universities. It seems that we are educating to be “directors of projects” as employees of great companies or in worst of cases government employee.
We are not transmitting the idea that they can also develop programmes that many people can use, that they have the technical basic knowledge to learn much more and to try new things. We don’t even teach them the “basic culture” that I consider an enthusiastic computer science professional must have.
How many engineers know Slashdot or Barrapunto? How many have participated in projects of open source software? Who have read the code that others have developed? How many have participated of one of many programming contests? How many read the essays or articles of Larry Wall (Perl genius), Paul Graham (gurú of start-ups), Joel Spolsky, Marc Andreessen (creative of Netscape) or Lawrence Lessig (creator of Creative Commons)? How many people read the blog of Jonathan Schwartz (CEO of Sun) rather than just talking about and demanding to be taught Java? How many consult or collaborate with Wikipedia? How many are eager to make programs that “people” use not “companies”?I can assure that a minority — professors too — know about this. It is a shame, but I believe that it has a lot to do with the motivation that IT professionals feel in the career that they have chosen.
I always say it in provocative plan. Yes, if we want to change the university it would be necessary to dismiss 50% of the professors –myself among them– but to be coherent we also would have to throw out 80% of the students.
In the end the bottom line is if we want elite universities like Stanford or MIT or “more social” universities. I undoubtlessly support the second option, as it has its unquestionable social advantages, but also it has its associated problems: shortage of [good] professors, lower wages, fewer resources per capita, overcrowed classrooms, students without the motivation of those of the “elite”, etc.
If we want to maintain this model we will face those difficult and complex problems, but it would be necessary to avoid demagogic simplifications. However it would be best if we attack the fundamental problem that does not require a great budget – how we motivate firstly the students and then the professors?
Perhaps if we minimise this problem, the rest that is needed to form entrepreneurs will come: to have nerve, self-confidence and to stop looking for the security of a great company. Someone with a title of engineer at age 24 and living with their parents has absolutely nothing to lose if they want to prove something new. If even a sad professor of operating systems in his forties, associated to a freelance programmer friend and programming in a miserable language as PHP is called for interviews, can try when they begin knowing infinitely more than I did at their age [*]? ; -)
[*] I am very serious. There is no relation between the programme that my students attend and the one I have done.
(1) SL: Limited Company in Spanish
(2) LOPD: Ley Orgánica de Protección de Datos. Spanish law to protect private data.
(3) LSSI: Ley de Servicios de la Sociedad de Información. Spanish law applied to services provided through electronic media.
(4) IVA: Impuesto al Valor Agregado. Spanish tax similar to GST or VAT in English speaking countries.
Stumble it! Bookmark it! Digg it! Menéame! Fresqui Suscribe Suscribe