Why Sondz?

 

A little under a year in (in its public form) and while many people around the world are enjoying a well-earned summer break (if and where they can), let us take a slight step back from the more day-to-day aspects of running and scaling a platform like Sondz. Instead, let us focus on the arguably equally interesting question of how it ever came to be. As Elon Musk once quipped (and as echoed by many), ideas are not the hardest thing to come by — it’s implementing them that is. With a small nuance: when we talk about implementation here, that also entails the more minute aspects of said original idea. In other words, we are going to talk about the journey from a global, top-level idea and how refining it made it stick — and eventually work. So far.

 

Finding an idea

Most great ideas are not all that great if you think about it. Meaning: when taken into the context they were developed in, they tend to make sense. CSI creator Anthony Zuiker famously stated that his original script broke all TV crime drama rules because he didn’t know them: rather, he formed the idea while watching a true crime forensic documentary and just rolled with it. Likewise, the pioneers behind Pyra Labs (who would go on to found Twitter, Medium and Square among others) came up with their first major breakthrough because they were trying to keep up with each other while working remotely. Blogger was born, and in many ways blogging as a concept too. All these years later, as you are reading this wonderful piece of prose, you are still benefiting from that eureka moment…

In our case, the original idea was fairly straightforward indeed: we couldn’t quite figure out why cinema had comprehensive databases when music did not (see IMDB, of course, as well as Rotten Tomatoes, Box Office Mojo, All Movie Guide…). We found out later there was a key reason for that: music is a far broader market, with many more actors, and therefore a far more complex topic to tackle. But, back then, our younger and more idealistic selves did not think about that. Only about the fact that the world’s premier art form did not quite have a comprehensive information platform of its own. Sure you could find great websites to browse second hand discs (Discogs), to read professional reviews (All Music Guide), to stream music (Spotify & al, of course), to watch videos (Youtube, also of course)… But none that was focused first and foremost on core musical data and aimed at a large audience of fans and amateurs alike. Hence, Sondz…

 

Refining an idea

Once we had that initial insight, then came the eminently important moment of fine-tuning that core idea. What does a “comprehensive music platform” look like? What are its features? Where do you find its contents? How do you name it? Granted, that last one was perhaps not our greatest challenge — the others were, though. The key dual question we had to answer was in effect: what do you want to show and where do you find it, both elements that are so deeply intertwined that it can potentially become a chicken and egg situation…

On question #1, i.e. structure, it took us a while to determine — and prioritize! — what we thought was key and what was less so. Our painstakingly long prototyping work included a lot of back and forth between everything we wanted to show, everything we forgot we wanted to show but added post facto, and how it could all look together. This resulted in dozens of iterations for our 4 key templates — artist, release, song and label — with the trickiest being artists. Naturally, that pages needs to incorporate a wealth of information as it meshes together personal information as well as band facts (if any), discography/release elements, songs, various media, external links… The “Facts” sections alone, which can include dozens of fields that differ depending on whether you are a solo artist or a band, took a while. Then you had to come up with all the wordings, placeholders… Fun.

On question #2, i.e. data, that one came with its own set of surprises. Our initial thought was that we would essentially scrap Wikipedia. After all, that platform is open, contains quite a bit of information relating to music, is fairly easy to operate… But there’s a huge caveat: because of its inherently flexible nature, no single Wikipedia page resembles the other. Which makes the platform so utterly adaptable to any topic — while our platform was the exact opposite: Sondz focused solely on music and we therefore worked on designing the best experience for what a musician’s page would specifically look like. It could be any musician, but it had to be a musician. As a result, integrating highly heterogenous content from Wikipedia onto our structured templates would have proven difficult. In the end, we did use some of Wikipedia’s information, namely page overviews, as these don’t necessarily need to be structured to fit our platform. We then found other more specific platforms to feed us the rest of the information — on a more structured basis…

Final comment we will make on this stage of the process: the fine-tuning phase also involved a significant slimming down of our MVP (Minimum Viable Product). Without going into great detail because we still very much intend on adding those various features in the not-too distant future, our initial concept was far more technically complex as it involved a lot of user interaction — i.e. the ability for users to directly contribute to the platform. While we decided against adding that layer of technicality into our currently live Version Alpha, we hope to deliver on it in the coming months. For a first set of user-generated features at least: there will be more later, we swear!

From the original idea to the end of that second iterative step, months (if not years) went by. This included functional prototyping, to be fair, but no actual end user coding. Which was enough for us to test out our assumptions on a product and user experience (UX) level, share our prototyping with trusted friends and acquaintances for feedback, but did not include public testing per se. That came later, as the greatest test to anyone’s concept you will ever find…

 

Testing an idea — live

No matter how much you fine-tune and pre-test an idea, nothing is more efficient than actually releasing it to its intended target. You may be the smartest developer around, your product may sound and/or look great, a single day live may have you rethink its very core. Luckily, that is not quite what happened to us, but these past few months spent observing user interaction with our platform, fixing certain features and introducing new ones we realized were lacking made all the difference.

Without blowing our own horns too much, it turns out all this time spent prototyping and (over)thinking information placement, placeholder wording and/or data management saved us from a day 1-type catastrophe. Not only did the website hold, but we didn’t get any horrible feedback such as: “How come this huge piece of data is missing?”. To be fair, we didn’t get a whole lot of applause either, which is OK by us: a satisfied customer is one for whom your product is intuitive and who will therefore use it so naturally that it does not necessarily register that they are to commend you for it. In other words, the initial launch went fine, our biggest and already publicized problems came later on the technical side of things, as we worked on growing our user base.

There were still quite a few things we overlooked if we’re being honest — and we very much try to be. For one, we started out with a Google-like homepage: nothing on it, save for a search bar. Which looked cool on our prototype, but made less sense while live: it turns out people favor suggestions when arriving on a website, Google very much being the exception to the rule… Same goes with search, which we also previously talked about: our still ongoing work on making search algorithms intuitive enough had some users abandoning their search when first results proved unsatisfactory. You can’t blame them for that: in 2021, people expect quick results, rightfully so.

Another classical topic, which forms the basis for much of the queries/questions we receive: Sondz does not show all of music. To that we will simply answer: that is absolutely correct. No one else does, or ever will: back to our original point here, the music market is simply too rich and dynamic for that. Which translates to some information missing, some artists missing, some wrong/outdated information being displayed… The issue is particularly prevalent when referring to more niche types of music: when we tested our then-young platform with the member of a leading classical music string quartet, he immediately found issue with information displayed on his work: a member of his quartet had actually left but was still mentioned as active, thus making the quartet a quintet… Although such situations will inevitably always occur once in a while, our goal is that they happen only once in a long while!

Full disclosure: we should have live tested earlier. In other words, we should probably have launched sooner, even if our product was not all set up, even if we were not all totally happy with it (if anyone can be totally happy with anything…). Agile theory is all about testing quickly and adjusting/pivoting if need be: we knew that theory but also realized it was quite hard to fight off our overly perfectionistic instincts. While we did bring ourselves to ultimately show our baby to the world, we would have likely saved some time doing that sooner, if only to receive some of the key feedback we got early, iterate then and there and move on to the next batch of features safe in the knowledge users will be on board… Live and learn, always!

 

The future of an idea

Again, Sondz Alpha is not even one year old — it can hardly walk. Which means 2 things:

1. We still do not have a track record long enough to tell what the longer term dynamics of the platform are. In other words, we cannot honestly state that we will take over the world just yet, even though we obviously think we will;
2. There are still a number of ideas and features we wish to add to our initial, overarching concept to make it fuller. Only then will we truly be able to tell if it “worked”. It won’t be all that hard to tell either: you will just have to listen to the users…