What tech can learn from the sorry saga of the Sonos app

Sonos’ app redesign has been one of the biggest tech disasters of the year so far and there are lessons in what happened for everyone.

I really like Sonos' products. They're well-made and generally work well. You've probably read about the company's recent troubled release of its new "app," a piece of software that sounds simple but has proven challenging to update for some users and flaky in operation for others. (If not, catch up with The Verge’s piece here. Ed) It has been a PR disaster for the company and has much more profound consequences than that. I think we can all learn from it.

I'm not one of the baying mob shouting for the CEO to resign or accusing Sonos of stupidity or dereliction of duty to its customers. But I can see how this happened, and with the best will in the world, the company should not have been led down this path.

Even if you're not interested in Sonos, the lessons here can apply to any technology.

Early days

I first became aware of Sonos some time in 2007. At the time, the wireless music landscape was essentially non-existent. Sonos has a detailed and refreshingly honest story of its formation and early days on its own site: https://www.sonos.com/en-gb/how-it-started

Here's a flavour of it:

"In 2002, great music in the home meant wires hidden behind bookshelves and furniture, connecting to speakers the size of bongo drums; audio jacks plugged into the right holes on the backs of receivers and players; physical media primarily in the forms of compact discs and tapes - and if you wanted a multi-room experience, an afternoon (or weekend) drilling through walls to snake wires from a central receiver to speakers throughout your home.

While the original Napster had risen and fallen as a means to find music online to play on the personal computer, digital music was still new, and the idea of streaming music directly from the internet was far-fetched. Pandora, iTunes, Spotify, and the rest of today's leaders in music streaming services did not exist, nor did the iPhone. The top Internet service provider in 2002 was still America Online via dial-up, and fewer than 16 million US households had high-speed broadband."

Beyond its original innovation of wireless whole-house music, Sonos has done incredibly well in embracing the modern situation:

• WiFi is almost universal.
• The internet has plenty of bandwidth, even for lossless music streaming.
• Voice assistants are part of the typical domestic scenery.

During Sonos' gestation period, dial-up internet was commonplace, and even proper always-on broadband was pretty slow. But its vision was ambitious and, with hindsight, entirely realistic.

Shifting sands

Sonos has developed what is probably the best hub or nexus for music from just about every source in the home. The problem is that it's built on moving foundations. What users see is an app; what the engineers see is a swirling sea of shifting services and standards, all working harmoniously together until it isn't.

Just to be clear, I haven't had any severe problems with the new app. It's a bit sluggish, and when it was first released, there were some bizarre omissions like a numerical scale for volume. Many of these have reappeared. But I've spent time on Sonos forums, and while some of the "complaints" lack nuance and understanding (as do the replies: "It's your WiFi!"), the only thing that's clear is that there are deep-seated problems with Sonos' new software.

So why have I escaped the most severe issues?

Probably because I have the very latest products. You can read about my Sonos Era 300s elsewhere in this publication. They're absolutely brilliant. But some Sonos customers go back a long way, and they don't use their systems the same way that I do. Some have enormous music libraries stored on hard disks. These rely on ancient media-sharing protocols that have existed for decades. New app users have struggled to connect with and manage their libraries.

Some users have as many as fifty Sonos devices, not to mention WiFi routers, extenders, and mesh network devices.

There's a very good article by Andy Pennell, a Principal Software Engineer at Microsoft. As a sideline, he's built his own Sonos apps almost since the inception of the wireless music company. It's a good read.

50/50 hindsight

So what should Sonos have done? (There's nothing worse than unqualified outsiders giving advice to troubled companies, but I have been a CTO, worked for a US PR agency, and was a Director of Communications for a software company.)

First and crucially, Sonos should not have released the software at this stage. But that is so obvious that there must be more to it than that. Here's what I think happened. When Sonos released its new app on May 7 2024, that date had been in the diary for several weeks. There were press previews, which were generally positive: the app was a little hard to navigate, but that was because it was different from the old one. There were few, if any, complaints about it being hard to set up or not working - I suspect because the chosen group of reviewers were tech-savvy, "friendly", and given unlimited, timely support to ensure that everything worked as it should. Crucially (this is what I would have done with a PR hat on) the reviewers either already had, or were loaned simple systems that were known to work well with the new app.

In my view, the primary issue was not technical but strategic. The company publicly committed to a release date before the software was finished. Imagine booking passengers onto a brand-new airliner that wasn't thoroughly tested. You just wouldn't. And you couldn't because there are regulations. But domestic HiFi is not exactly life or death, so those rules don't apply.

But, nevertheless, software isn't finished until it's finished. There have been plenty of other examples of software being pre-announced and not ready. You probably have your own favourite examples. Typically, you'll download a new app - or an upgrade for a DAW or NLE - and you'll find that perhaps your audio interface doesn't work anymore or that your Bluetooth headphones no longer automatically connect.

We're kind of used to that scenario. With computers, you can poke around and often fix these issues with a download. With Sonos, it's different because it is a closed system. There are, effectively, no serviceable parts inside.

A complex MVP

When software companies plan new products, there's a concept called Minimum Viable Product (MVP). It is an early version of the product that has the least amount of functionality that makes it still worthwhile as a product. Essentially, it's the most basic product that people will still want to buy and will find (minimally) satisfying. This approach has a lot to recommend it. Without all the complexity of a complete product, it's easier to test, easier for customers to understand (although its user interface might be immature) and it will be easier to diagnose and fix if there are problems.

I'm sure, in some sense, Sonos' app release was an MVP, but unfortunately for them, even the most minimal viable product is an extremely complex one. Here's why.

Sonos has an incredibly complex software and hardware ecosystem. The Era 300 has six loudspeaker drivers, six amplifiers, powerful digital audio processing, and Ethernet and Bluetooth. It has to work with Apple Airplay and handshake with an increasing number of streaming music services, internet radio stations and third-party speech assistants (like Alexa). The app has to deal with wildly different network topologies. Sonos' software has, in addition to its app, cloud-based components. It has to deal with latency - the almost inevitable delay when you request something from the cloud. If the old system needed replacing because it's a mess, it's tempting to think the new system also needs replacing because it's a mess.

Forcing its hand

You have to admire Sonos' ambition. You can understand that the company needed to update its app.

So, why didn't Sonos wait until the app was finished? It may have felt that having announced the new software so confidently and publicly, it couldn't delay it. Obviously, that would have been embarrassing but nowhere near as painful as the current situation.

But there's another reason: there are more products to come. There are rumours of an Era 500. The new Sonos Ace headphones need the latest app for full operation (as opposed to basic Bluetooth operation, which doesn't need the app).

The bottom line is that Sonos needed to release its app to allow it to release its new products. It has somehow manoeuvred itself into an impossible position.

It will now have to focus on the app. It will have to delay its product releases, which means that its revenue will suffer. It has recently had to let around a hundred people go. Some potential buyers will be deterred by the current situation.

But I hope it pulls through. This could have been managed better, but nobody's perfect.

tl;dr

• Sonos' recent app redesign has turned into a major PR disaster and has had significant consequences for the company.
• The new app has encountered deep-seated problems, especially for users with older Sonos products and extensive music libraries.
• Lessons for the rest of the industry watching on include not releasing software auntil its ready and conducting thorough testing to account for various user setups and scenarios.