Songsterr - the results of three years

Friends and colleagues!

In this article we would like to share the experience of our work on the site Songsterr.com and talk about the results achieved. We hope that our experience will be useful to other developers.
')
About the project . In short, Songsterr is a catalog of guitar tabs with the ability to play them through a Flash player. The user not only sees what notes need to be played, but also can listen to how this or that piece to be learned should sound, which is very convenient. We will not go into details, anyone can see the capabilities of the player, for example, via this link .



Key project indicators:

• over 110,000 users daily
• monthly audience over 2 million
• more than 70,000 likes on facebook

Prehistory More than three years have passed since in February 2008 we published our first post about Songsterr on Habré . Having collected feedback, at the end of July 2008 we released a new release , which generally formed the concept of the project. The concept turned out to be working, and we can say that Songsterr today is a direct continuation of the July 2008 release. In August 2008, Techcrunch , Mashable and Lifehacker wrote about us; Songsterr was recognized as the best startup in August on Habré.

What's next?

It was clear how to develop the player - at first it supported only acoustic songs with one track, then support for various instruments and songs with several tracks was added, the ability to slow down the track, print, full-screen mode and other functions. Actually this plan was voiced in that July 2008 publication. It was also clear that in order to attract a wide audience, a large number of songs are needed. For users, the opportunity was created to upload new songs and update existing ones. To speed up the process and improve the quality of the content, we entered into a dialogue with the most active contributors, motivating them to add more songs, and also attracted guitar teachers who could evaluate the quality of the content and act as moderators.

Monetization . What was completely incomprehensible was how to monetize the project. In the fall of 2008, we had about 1,000 visitors per day, the project was completely free, the only source of income was AdSense advertising. The option to wait with monetization or even not to think about it at all did not fit, since we had neither an external investor, nor large own savings that would allow us not to think about money for a while. In addition, there was an understanding that sooner or later this question would still have to be resolved and better resolved before, since attempts to take money for what was free would inevitably cause a wave of negative users. If, from the very beginning, we distinguish between paid and free functions, this is perceived as normal.

Some ideas of monetization were placed in a separate post on Habré, and even an interesting discussion ensued in the comments. We did not like the idea of ​​taking money for access to the catalog, because the catalog was not filled by us, but by users, and that would be unfair. In the end, we decided to try the Fremium model in relation to the player functionality. The basic functions of the player remained free, users receive access to the “advanced” functions of the player by subscription, for a fixed amount per month - $ 9.90. We called this advanced plan Songsterr Plus, its comparison with the free plan is available here .

In general, the Fremium model is not as simple as it seems. Some comments and difficulties encountered:

• The free version should remain comfortable enough to continue to be used. The free version creates the basis for further growth. Thanks to her, search engine traffic is growing, users tell their friends about the site, put links, do tweets.
• At the same time, there should be restrictions in the free version, otherwise no one wants to update to the paid plan. About a year and a half ago, we tried an experiment - for free users, the ability to play tabs was limited to the first 30 seconds. Yes, a 15% increase in subscription conversion was obtained, but at the same time the growth of traffic completely stopped, and we received so many angry letters and comments that we decided to suspend the experiment. The restrictions were too tight.
• price selection. In general, it is not very inconvenient to carry out experiments with the price, especially if the project is at an early stage. We have to wait long enough to understand the effect of the changes and collect statistically reliable data.

We began to receive the first subscribers, but the conversion was quite low. The good news was that on average, paid users use our service for at least six months. A few months later, we had already gained a critical mass of subscribers, which yielded income several times higher than advertising revenue. In general, the model with a subscription is very well calculated, and there are plans to write about this separate post.

Revenues from advertising . There is an opinion that AdSense is only screwed onto the most lazy webmasters on your website and there are ways to get more (or even an order of magnitude) more from the same sites. Perhaps this is so for individual areas, but frankly, we could not find anything radically more profitable than AdSense.

We tried other networks — for example, Valueclick; we tried selling guitars and guitar equipment through our partnership programs Amazon, Shareasale, and Commission Junction. Every two weeks we receive a letter from the next advertising agency, which offers us to run advertising for their clients on 50/50 conditions promising fabulous profits, but in reality it turns out to be even more unprofitable than AdSense. The only exception is that the partnership with guitar sites, video lessons, gave a good result.

In general, there were no large bets on advertising revenue, so, for example, we didn’t try to use advertising modules to cover all free space on the page and didn’t use aggressive advertising.

iPhone / iPad application. In March of this year, the Songsterr iPhone application was launched, in April we released an update that added support for the iPad. Currently, the application costs $ 4.99 and takes about 40th place in the US top paid applications category "Music" for the iPhone and about 15th place in the same rating for the iPad.

The application in April fell into first place in various European countries and stayed for about a week in the Top 10 in the United States. The effects associated with getting to the top and our experiments with the price are the subject of a separate publication.

Facebook Since January 2011, we stopped updating the company's blog and completely switched to the facebook page . If the news in a blog in a rare case received 1-2 comments (usually 0), then the facebook response is much higher and faster - these are dozens of comments, the first of which appear within a few minutes after the post. Users vote "likes" - it is easy to assess what is claimed by the audience and what is not.

Also, some time ago, we abandoned our commenting functionality and switched to the comment system from facebook, which also had a positive effect. There are fewer comments requiring moderation; users have begun to help each other more actively, for example, to respond more frequently to questions from other users.

Team Currently, the project is supported and developed by 6 people: project manager, manager, designer, three programmers.

From the very beginning, we were inspired by the ideas of 37signals, so the working tools of the project are Basecamp and Campfire.

Immediate plans .

• site development;
• new functionality online guitar player tabs;
• development and launch of an Android application;
• iOS application development

By the way, in connection with all this, we need an employee .

If you have questions related to the topic, we will be happy to answer.

PS Of course, the true recognition for us was the mention on lurkmore :)