This report is over a year late -- mea culpa. 2014 and 2015 were challenging years for MetaBrainz where we struggled with wanting to do too many things, while still being severely resource constrained. Many things fell between the cracks as a result of this. However, now that our efforts from 2015 are bearing fruit, we have more capacity overall to focus on catching up on all the things that remain incomplete from the past few years.
In January we completed the hand-off of duties from Oliver Charles to Michael Wiencek (Bitmap) to fully bring Bitmap on board, Nikki, Ian, Bitmap and myself met up in Chicago to do some planning for the future. We rented a dingy AirBnB apartment in Chicago in the middle of the polar vortex. -23C is a great temperature for keeping everyone indoors and focused on the tasks at hand!
In February we officially deprecated the embedded RDFa in our pages, only to have Google come and ask us to add JSON-LD markup to our pages two weeks later. In March we started planning for our upcoming schema change in May, setting ourselves off on a less than ideal pattern of having too many schema change features. We would put too many features on our plate and then had to spend too much time cleaning up the schema change release. Sometimes this process would take us all the way to the next schema change process.
Our schema work would take large chunks of April when we also participated in the next round of Summer of Code. In 2014 we accepted projects to finish CritiqueBrainz ( https://critiquebrainz.org ), start work on a new SOLR based search infrastructure, MusicBrainz UI improvements and a new site for Picard.
In May we carried out our schema change and amazingly we released CritiqueBrainz at the beginning of Summer of Code! This is when Roman Tsukanov (Gentlecat) became an important contributor -- he carried out most of the tasks he set out to do over the course of the summer before summer even got underway. We even cooperated with Google on a blog post about our success.
The summer was quieter for us as we finished up the last bits of the schema change release. This quiet period showed the first visible signs some members of our team were hiding from the public to carry out their duties. Many discussions took place behind the scenes to try and resolve these issues, with very little progress being made.
In September many members of the core team met up in Copenhagen for our annual summit. This is where we met Laurent Monin (zas), a long time Picard contributor, for the first time. While we had a lot of fun during this summit (we had rented a family home complete with back-yard) team divisions became more clear.
At the summit we also agreed to reform the style process and to appoint Nicolás Tamargo (reosarevok) as our style BDFL in order to try and streamline the style process. We also started preparing for the fall schema change, which in 2014 was scheduled to happen in November.
Besides working on the schema change, both Alastair Porter (alastairp) and Robert Kaye (ruaok) started working the AcousticBrainz project which aims to collect acoustic information about music. The project was soft launched after a massive 2-day hacking sprint.
The rest of the year was mostly spent working on finishing the bumpy schema change release and getting AcousticBrainz ready for a wider audience. In December Robert Kaye met with a pro-bono brand consultant whose goal was to guide us on our journey to create a consistent brand for all of our projects.
Finally, we should mention the Geordi project that aimed to be a repository for 3rd party music metadata provided to us from partners. We worked on this project on and off in 2014, but the overall goal remained elusive and we never really made much progress on the project at all.
As part of our situation in 2014 where we struggled for resources to even keep existing infrastructure running, we had our traffic analysis system break as it was outdated and overloaded by our traffic logs. This proved to be a significant project to fix and even in 2016 we have not addressed this. This means that we cannot provide an interesting traffic graph that shows how our traffic changed over the course of the year. We aim to have these traffic stats return at some point.
Our traffic can be broadly described as being steady over the course of 2014. We would routinely get 20M requests per day and we were able to honor about 16M of those requests. The remainder of the requests we could not honor (and resulted in a 503 error) due to bandwidth and hosting capacity limitations.
The need for a new hosting solution became clear in 2014 and we started examining the various options for moving to a new hosting provider.
Our finances in 2014 are summarized by our Profit & Loss statement:
|Web Service Access||420.00|
|Donations - PayPal||5,776.82|
|Donations - General||55,874.04|
|Consulting - General||2,500.00|
|Live Data Feed||171,210.00|
|Bank - Interest||60.53|
|Affiliate - Amazon||989.71|
|Affiliate - Tagger||14,213.23|
|Licenses - CC data||6,000.00|
|Affiliate - CD baby||11.50|
|Hosting - Hetzner||420.49|
|Development - 1099||75,760.38|
|Bank - WePay||53.65|
|Bank - General||745.31|
|Bank - PayPal||1,029.03|
|Rent - Higuera Office||2,400.00|
|Development - Contractors||1,231.62|
|Hosting - DWNI||20,700.00|
After a turbulent 2013 we managed to keep our finances on an even keel this year. Income was fairly steady and most of our customers paid on time, so we didn’t see any payment backlog like we did in 2013. While 2014 wasn’t as good for us as 2013, we did have an excess income of $31,164.73 on $257,055.83 of income.
In the course of 2014 we spent $1,087.15 on hardware and $20,700.00 on hosting costs for a total of $21,787.15. This compares to $24,756.65 in 2013, which means that our costs were a little less in 2014. Sadly, we will not be able to compute the cost per 1M web or web service requests until we get our traffic stats system running again. In 2014 we spent $86,992 in salaries for developers/community management and our costs for administration was $79,999.26.
We earned $171,210.00 from our Live Data Feed and $6,000.00 from Creative Commons licensed data for a total of $177,210.00, which is a 6.5% decrease over 2013. Towards the end of 2013 AOL Music stopped operating and in 2014 Intertech Media stopped using our data and Zvooq stopped paying us. However, the BBC decided to pay 3 years ahead on their Live Data Feed costs, which made up for most of the other lost customers. Overall, our Live Data Feed income decreased by $8300.55 in 2014, which is effectively flat for the year. Many thanks to the BBC for saving our income stream! Our web service access income in 2013 was $660.00 and only in $420.00 in 2014, which suggests that charging for our web service may not work as well as we had hoped.
Our end user donations via PayPal came to $6,035.07, which is up 4.5% from 2013, but still far less than 2012. Direct/large donations added up to $55,874.04, including $40,000 from Google for our annual support and a $10,000 non tax deductible donation to allow us to utilize Google Cloud services. Also, Spotify once again supported our summit with a $3,000 donation.
In 2014, 19,397 editors made a total of 5,234,229 edits. The total number of active editors is up from 18,710 in 2013, and the total number of edits has increased by 186,622 edits.
In 2014 we received 1,235,995 edit votes, which is up 230,049 votes from 2013. These votes came from 1187 editors, down from 1243 editors in 2013.
At the end of 2014, MusicBrainz had 20 machines in service, largely unchanged from 2013. From the top, going down:
MusicBrainz uses 15Mbps of bandwidth per second on average and draws 35 Amps of current for a power consumption of about 3,850 Watts. MusicBrainz physically occupies 20 Us of space (half of a rack) at Digital West in San Luis Obispo, CA.
Many thanks to our editors, voters, peer reviewers, bug watchers and other members of our community -- without you MusicBrainz would not be what it is today!
We'd like to also thank our developers that pushed out dozens of releases of the site, Picard and our numerous client libraries. All of your work is critical to enabling the MusicBrainz community to do its job.
In particular we'd like to thank Google and Spotify for their large donations made in 2014. We’d like to also thank the BBC for pre-paying its license fees! These large contributions/payments allow us to carry on with our goal of making MusicBrainz the most comprehensive music encyclopedia out there!
Thank you to everyone who contributed in 2014!