Raise the software business

This is the second article about the development of a software (Internet) business in Russia (the first one is here ). In it, I am going to reveal some key points of adapting a product to the market (according to user requirements), about overcoming the tenological gap, and a little bit about the realities of Russian business.

This topic I want to coincide with the fact that a year ago (maybe a little earlier or a little later ) appeared the product WEBO Site SpeedUp (formerly Web Optimizer). Thoughts are a bit messy, but, apparently, for each direction everything is very specific, therefore only general ideas, directions and approaches to solving problems will be valuable (which I sincerely hope :). Hereinafter we are talking about creating a box solution (and not software outsourcing).

First, a few quick answers and refutations.

1. Why do you need a legal entity for business?

As it turned out (I personally had no doubt, but there were different opinions in the comments) it is much more convenient to sell software with just a jur. face. This and a smaller amount of taxes (in aggregate is now 6% from the shaft + 9% from net profit / dividends versus 27% for an individual or 45% altogether including VAT). This and greater trust on the part of potential customers (yes, the contract to enter into with SP is still more dumb than with an LLC with a legal address), and there were contracts (but you need to go here on sales volumes). This is also a normal opportunity to work with foreign money. However, if you do not plan to rise above a turnover of 100k rubles per month, then the legal entity will definitely be a burden.
')

2. Why do we need a "fake" jur. address?

Its main advantage is cost savings. For example, half a year of rent under this scheme costs 7-10 thousand rubles. If you issue all the rank according to the rank (at least with one workplace somewhere in the co-working center), then it will cost approximately 4 * 6 + 15 thousand, i.e. 4-5 times more expensive (maybe you know cheaper alternatives - welcome in the comments). But if you reduce costs, then it is better to wait for a normal office to take off until you have 2-5 employees on staff.

The only (more precisely, there are two of them) minus from such a deal - they do not open an account in a “decent” bank (this is Alfa-Bank + m. Raffaisen and others like them, VTB opens it with a bang). + There are some costs for correspondence, but if you do not work with the whole world under contracts, then this is not about you.

3. Why do we need a normal merchant?

(In this context, this is a processing center for payments for your products, in fact, the keys to software products). Why not just "take cash"? First, (unless you have a product that costs 5,000 rubles or less - although in this case you will hardly pay such an amount, more under the contract that the same bells are just on the side), you will spend valuable time writing out the keys each order. And secondly (with a rather high probability), every second buyer will be personally with you (of course, the whole process can be automated, but you calculate the cost of automation, and you will understand that at the above revs it is much more economical to “surrender” to the merchant) to ask where to enter the key and how to use the product. This time, time and time again. If you, of course, do not sell technical support along with the product itself, then here the business will be very much worth it.

A few words about Softkey, Allsoft and some other options. Soft key is better (less commission) + there is an opportunity to work with foreign buyers (via Softkey.net). The withdrawal of money goes immediately to the account (without additional checks, commissions, etc.). Allsoft on its background seems to be some kind of box (all persistent requests to come to the office in person and sign the contract were rejected - the papers were only sent by mail), their commission is, on average, higher. But (IMHO) more powerful marketing. What and need to use. Ideally, of course, it is most advantageous to have your own processing (for example, based on the same insales.ru), and through these stores to attract extra. clients (and, maybe, to maintain an affiliate network). More specific information yet. Maybe in six months there will be calculations with live numbers.

Personally, I have requests on Skype, “Let me transfer 500 WMR + 30 WMZ + money to the phone - you will sell the product to me, and at the same time you will explain how to set it up” already tired a little :) Therefore, I’m a normal merchant (Softkey for some reason more in terms of experience working with international-level stores - I’m not at all satisfied with it, but there’s definitely less of them all evil).

Yes, a small remark: hereinafter we are talking about building the most "white" business, and not about creating the button "to cut the loot" :)

From the lyrics to the case

Some questions have been clarified, now we are moving to the most interesting part - what will work for the development of a business (or a product - like someone more like it, a software business without a product is the same as March 8 without flowers and sweets), and that will not work particularly. In principle, something new in the following facts is unlikely to occur (all this has already sounded more than once, and not two), but, perhaps, the aggregated experience will help someone.

Placement in profile catalogs

Firstly, if you go from the needs of users, then you need to go to the target groups and communicate with them (to offer a product, evaluate response opinions, make improvements). For web products (consumer level) there are not very many options here - these are the communities of the most common CMS - WordPress, Joomla !, Drupal (and smaller / local).

I will dwell on the experience of communicating with 4 directories - these are extensions of WordPress , Joomla! Extensions Directory , Drupal and Bitrix partner modules . Oddly enough, Drupal turned out to be the hardest place - the moderators strongly rejected all attempts to add “something non-standard” and “not conforming to the Drupal coding style”. Due to the small number of the Drupal community, attempts to add there were temporarily stopped.

WordPress was the most pleasant place to create extensions. Extensive API, complete freedom + no moderation (only reviews of the users themselves). As a result, it took about 3 days from getting acquainted with the API to publishing the extension (for example, for Joomla! It was 2.5 months, for Bitrix - about a month). When developing for WordPress, you get access to the shared SVN, while all updates are synchronized with the site, the readme file is translated into a set of web pages, and all users receive a notification about the update of the extension. Lepo, and only.

For Joomla! (Apparently, due to the lack of moderator resources) the initial consideration of the application for publication in the catalog took a month (if there were kicking once every three to five days, it could take three months) + another month went on to reconsider the rejected application.

For the Bitrix API, it is somewhat confusing (probably due to the payment system and the lack of a sufficient number of ready-made extensions for review), but the task is also solvable.

According to the results of WordPress traffic with Joomla! give about equally, Bitrix with Drupal also about equally, every 5 less than the first two. It speaks, more likely, about the volume of the community and the popularity of the systems (and also, about the expediency of developing them).

Also quite strong (worldwide) is the MODx and Magento community. If you develop something - focus on them too.

Technological abyss

I can be mistaken in the exact term (if someone corrects in the comments - I would be grateful), but here I mean a significant gap between the innovative product (for geeks) and the mass product (for ordinary users). These two categories differ significantly, and if you work for the mass market (and start by creating a new product), you will inevitably run into this chasm.

It is worth considering that the same "product for geeks" and "mass product", most likely, have completely different characteristics. Geeks will appreciate the number of functions, potential, uniqueness. The inhabitants, rather, will be attracted by brand awareness, convenience, reliability. Those. If to attract geeks you can roll out a raw version that will regularly bash, break, but be “super cool” in capabilities, then for the mass market you need a mega-stable, stylish and convenient product that may not have all the “first-class” capabilities. Therefore, in fact, you need to “degrade” the product (no matter how sad it sounds) to a sufficient degree that it could be used by the “spherical inhabitant in a vacuum”.

In overcoming this gap, one should not lose sight of the main purpose of the product and make it massive, without losing appeal to the most advanced part of the society. This is the only way to be the first: otherwise geeks may be disappointed, or someone else will release a more “tricked” software product. In any case, recognition by the geek community should be the starting point for creating a normal, high-quality (from the point of view of the average person, of course) product.

This task is achieved relatively simply (not in terms of time or man-hours, but in terms of directions and ideas): you need to constantly collect user feedback, iteratively refine the product and not lose sight of the main idea (the one that was at the very beginning ). And sooner or later it will happen - you will receive a product recognized by the market and being the leader.

About time and limitations

If the development of a prototype of a product can take one to two weeks (or, as it is now fashionable to practice, one weekend), its output at least to the RTM version (Ready To Market, ready for the market) is at least a couple more months of stubborn labor. At the same time, it may take years at all to overcome this technological gap. Therefore, starting something to create, be prepared for a very large investment of time and patience, before you start to get something really worthwhile. And it would seem that a “simple idea” will take from several months to several years before becoming a “working product”.

Due to the extremely limited time and resources in adapting to the market, the only working rule is Pareto's law, 20/80. Each time you encounter or formulate a task, evaluate its priority. And remember that solving only 20% of the highest priority tasks will 80% advance you towards your goal, so don’t scatter on trifles (yes, it’s still important to learn how to evaluate trivia and what not, but it comes with experience :)

Sometimes it may seem that days and months are passing, but nothing happens - this is normal. Usually, changes accumulate, and only then there is a qualitative leap: you are shown on television, they write about you on digg, or the president presses you. But in any situation, we need systematic work, adaptation to the market and solving the tasks on the list.

A little more on the subject of limitations: the creation of a mass product has its advantages. You can collect a sufficiently large number of similar comments and cluster them, making up documentation, answers to questions, manuals, and so on. Naturally, this will only cover the majority of users (those who ask the same questions), but after spending a little time, you can satisfy them all.

The following orientation can help here: each time, when setting another task, think about what percentage of users will be happy after solving it. Enough? Then we must urgently decide. One or two? Then it is worth waiting until better times - maybe the task itself is reformulated or will disappear due to other changes.

Business model

There is nothing definite about this topic yet. If about 10 years ago the sale of software products worked out of the box, then now everything was strongly socialized, and the business models somewhat “floated”. At the moment, we have already changed 2 concepts of monetization of the idea (that is, the third is now in the process of testing). Maybe in the next topic there will be more information.

But one thing is certain now: for niche products (not the Linux kernel, not the DBMS or the web server), the open software model + donations work worse than selling commercial licenses: insufficient user coverage. Although everywhere you need some kind of line of features: this functionality is available for free, and advanced - for a certain price.

It is also obvious that the advertising model for software products is very limited (i.e., the browser or the site management system is still far away, but for extensions the browser itself or the system no longer exists, but I would appreciate it if someone gave the link in the comments on the enumeration of models of monetization).

UPD thought that I wanted to add: “bending” under the market should in no case forget the original idea. Otherwise, it turns out that “surviving at the beginning of the journey” you will be selling “high-tech milk” in a couple of years, and not doing what you started with.

Some statistics

I do not know how specific the figures will be interesting, but they can help someone for a general assessment of the effectiveness of development. So, over the past year

• 50k product downloads
• Every month an increase in downloads of 20-50%
• 1000+ working sites on it

At the same time every day

• 100k + Internet users load websites faster (sometimes without even knowing it),
• while saving about a million seconds.

The formula here is quite simple: loading the site takes 5-10 seconds, the average acceleration: 2-3 times. Each user views, on average, 3 pages on the site (already taking into account the positive effect of fast loading).

PS Yes, in the next issues I will try to illuminate more deeply office problems, salary projects and some financial subtleties.

PPS site www.webogroup.com is located on GAE (do not count for advertising, it's just quite convenient and, most importantly, free), so for the habr-effect we are now calm :)