Kubaka. Two years from the life of a lemur. Year 2012: Immersion

Hi, Habr.

Two years ago, two developers from St. Petersburg started making a service that solves a trivial, but quite relevant problem: monitoring JavaScript errors in production. What came of it and whether it is easy for a programmer to become a start-up will try to tell in our story. About mentalities, two long trips to the states, start-ups, introversions, victories and a copper basin, how much money is $ 200K, why the iPad is needed more often than a laptop and how to get a year of free hosting.

Let's get acquainted, we have something to tell.

')

Where are you, path, led me

We started to do the service, not knowing what a startup is. It was such a dream of every programmer - to make something of their own. Therefore, we will tell our story through the eyes of ordinary developers. We had the opportunity to start doing our project, and about what nerves every day of our own company’s work and what the entrepreneur most often doesn’t have time for programming, we learned a little later. In the meantime ...



In the meantime, we enjoyed the code. It may seem that the error log is very simple, but due to the characteristics of the front-end, javascript tracking is much more complicated than the backend, because each browser reports errors, with a different set of data, and IE translates errors into different languages . As well as minification, inline scripts, several static servers, that's all. And the responsibility is huge, because an error in the client script can break user sites, and a bad architecture can slow down. But we already had enough experience in developing high-load services, so for all the time since the launch of open beta, we have never caused any problems (except that the page loading wheel initially spun longer than usual). The architecture was divided into 4 components: the script is loaded on client machines and sends data to us, the aggregator receives error reports, filters them first and puts them in a queue, the processor parses the queue, runs it through analytics and stores the result in the database, the web interface allows it all watch. Technologies are not at all fashionable (MySQL, Redis, RabbitMQ, PHP, Java, jQuery, Netty, Nginx), but convenient for a very quick start, it was important for us to make a prototype, the first beta. Hosting Selectel in the cloud. The first logo was drawn by a girl from one of the founders. The whole site was made in gray tones, because they did not know how to work normally with color. Even the graphics engine wrote their own, because Google Charts had been drawing for a very long time. True, they still used it for more complex graphs and charts. Paid services like AmCharts did not fit, licenses for SaaS are too expensive. Charts from Google are loaded so long that the whole service was transferred to ajax because of it. Let the google statics be loaded at the very beginning, which will be for a second and a half (there is no caching) to be downloaded at each transition, plus the initialization of the graphs themselves. At the start there were already several competitors, but they almost did not show activity. For a startup, even good: any investor is experiencing, if there are no competitors. In the summer, they launched the first beta, let in users, fixed bugs, and here (this is a surprise) ran out of money.

I wish you, brother, was sitting at home

In the fall, a team of two coders who think that all startups are made up of yesterday's programmers first came to the world for the first event. Before that, we sat in the kitchen for six months and coded, registered an LLC (with the help of Elba and MoeDelo), searched for beta users, and so on. And then it began.



The very first event attracted so much attention to us that it was the hardest month in life (it was worse after that, but by that time they got used to it). Some presentations, funds, accelerators, incubators, business angels, everyone is writing letters and urgently want an answer, making a decision, deadlines for participation in some programs are burning (in September the “season” begins) ... let me go back to the kitchen to write code! But investments were needed, it was necessary to continue. There is also no experience, but everyone wants a percentage of the company. We decided to spend a month on parties - participated in several events in a row to figure out how things work out, to ask people for advice. In the end, we decided to take a chance and went to the accelerator in the USA, at least see the world (we didn’t see it, there was no time), at the same time they gave a little money, you can stretch for another half a year (together without marketing and infrastructure costs). For the next 3.5 months, we were engaged in some kind of garbage, to which we were not yet psychologically ready and did not know how to properly use all this (mentors, events, Silicon Valley around). But these months were decisive in the change of consciousness. The code itself is not needed by anyone. It is important how they use it, how it is found and how to find those who find it. And for this (sadly for introverted programmers), it is necessary to communicate with people live. Lots of.

First coworking smoothie

We started to suspect something. Programming has become less and less time, “some garbage” took more and more. Later it turned out that garbage is called business development (for a long time they tried to understand what this term means, although everyone around us said that we should do just that). Later, even work on marketing email was perceived by us with joy, as work on something directly related to the product. However, we still tried and released a lot of new features for the new year, creating a gap from competitors (started following users clicks, made all sorts of filters, learned how to pull out the stacktrace, show how many users were affected by the problem, in general, we were not ashamed to tell) . They registered an American company (accelerator condition), opened an account with Silicon Valley Bank and moved to an American hosting (ping, stability). The first money was spent on the qbaka.com domain (a thousand dollars, normal for a 5 letter letter, I had to bargain for a long time), before that it was net. It was possible to find a team that made an awesome video of the western level in less than a thousand dollars (now the guys take much more, but still cheap, write if you're interested). Seriously, almost none of the major services at that time had a level of clips that way. Found a guy who painted a new logo.

- Listen, let's hang in coworking, I have an idea for a startup, I have already done a test. With me smoothies!
- So, immediately went to ***.
(folk anecdote)

And yet the users themselves will not come, and there was really no money for advertising, even taking into account the pre-seed investment. We need partnerships, promotion channels (more on this in the next post). In terms of finance, we were generally the most economical startup team, saved on everything (buying a domain is an exception), although the fashionable model of start-up development now teaches you to be a drug addict: spend fast, die young. As it turned out later, though this is an exception, the approach turned out to be correct: with our project to spend quickly it is a true and meaningless death, and the savings allowed us to ensure slow, but rational growth, first to end up with a cash prize in the competition, and then to invest, start accepting payments. After all, the service was still free, we did not pay salaries, and the credit card almost reached the limit. All because we started to make the service simple and congenial - for ourselves, for programmers, and the market turned out to be one of the most complex in the market: global B2B. Companies will buy our service, not programmers out of pocket, because it is important for business to provide a service without errors, and home projects can be monitored and placed in a free limit. And selling to companies is not at all as easy as watching the virus growth of some social service. Budgets are arranged differently, different people make decisions, someone outsource the development altogether, while others are too large and mired in internal bureaucracy, some have a clean web service, and others have an offline business (we use the pizza delivery website, for example ).



One of the few sincere joys of a startup is to go through user sites, wonder how different they are (countries, business profiles, scale). From a social network for cats to a tourist resort in Canada, from a major media site to a dating site for cheating spouses, from Citrix to home pages. But this is at the top of the iceberg, and inside it is a hell of constant nerves, zebras of critical (within the company) situations. They are pressed on all fronts: deadlines for events and presentations, reports to investors, pitches to other investors, answers to users, meetings with potential partners, reports to the tax office, verification of documents, everything needs to be prepared and constantly everything goes wrong. That electronic key will burn on the flash drive at the most inappropriate moment, when the transfer of money between banks is hanging and if it is not confirmed quickly, the state will take them for itself. For almost two months, the accelerator turned out to be without money for us (as in that joke: a start-up comes to an investor, and an investor is also a start-up), although we have already arrived in the states and the last personal reserves have ended.

Hello, and the rain does not pass, but I am not dressed for the weather since the morning

Branil of Homer, Theocritus;
But read Adam Smith
And there was a deep economy
(Pushkin, Eugene Onegin)

If your team is small (advice: a powerful diverse team needs to be built from the very beginning, if we have resources for this, we didn’t have it), then in a startup you will have to be an incompetent specialist in all professions in which large companies are engaged in certain people. Programmer, lawyer, evangelist, accountant, marketer, copywriter, technical support, designer, director ... You need to understand the dilution of stocks, convertible loans, nginx settings, which are conferences on your topic, the convenience of the registration form, how much money to send to the pension fund and volume in what time is better to publish news. Do not just understand, and do it all. Every day: google about actions, read kilometer contracts, write job texts, invent video script, work out each scene in detail, come up with improvements to the next version of the logo, print a banner for a demo stand at the event, create a design for it yourself, give presentations, call up with the bank, to go to the post office to send documents, respond to twitter questions about the service, follow the updates of competitors, the list is endless. Oh, and when to program? At night, if there are forces after this important event where there will be many investors. And tomorrow is a meeting with a representative of Opera. And the day after tomorrow, mitap for frontenders where you can pop service. How easy it turned out to go to California for almost 4 months and see nothing at all but the Grand Canyon.



But the year was not bad. On the last day before the end of submission of applications, we submitted our startup to Web Ready. Then we didn’t know what we were going to go to America, so we got to the finals already being in Menlo Park, half an hour on foot from Facebook headquarters, in the same house with teams of Greek and Lithuanian start-ups. By car, of course, also saved, so often went half an hour to the nearest station Caltrain (train, which connects the Valley and San Francisco). It is not true that in the Valley it is impossible without a car, it is very pleasant to walk in the fresh air. It's a shame to refuse the Web Ready final, so we agreed with two good friends that they would represent our startup for us. For two weeks from scratch on Skype, we have taught one of them to speak with a presentation of our project for several calls. In the morning, as soon as we woke up, we turned on the live broadcast from the final. The difference with Russia is exactly 12 hours, at our 7 am in St. Petersburg just announced the winners. We won. 1 place. 500 thousand rubles (before taxes). It was the best moment in our history, there was no such joy even from the investments received later. Our friends helped us win, for which we received a generous thank you from the prize (was it a joke, to get into a start-up engine for even one day), along with the girl who painted the first logo. The balance was divided among themselves, it allowed to close the loans and stretch for another six months, critical before receiving investments. We still save at the office - we work in the kitchen, only in January 2014 we will move to coworking or office. There was also a mini-grant from Skolkovo for 5 million rubles, but it seems that we do not get it.



About the adventures in America, the launch of payment acceptance, accelerators, a major partner, lawyers and investments I will write in the next post, here is already a lot. If you came here to find out about free hosting: startups in the framework of accelerators and a number of other programs can receive free of charge from $ 1000 to $ 2000 per month for a year from the largest hosting companies in the US: Rackspace, SoftLayer, Amazon, Google and Microsoft. From December 2012 to today, Kubaki hosting was free, despite several servers with generously allocated resources and traffic of tens of millions of reports per day. I will write more about this in the second part of the story.

PS We are looking for a talented front fender with a developed sense of style. If you are, write to amima@qbaka.com