In Third Republic’s DevOps Leaders Q&A series we speak with industry leaders to hear their thoughts on the ever-evolving technology landscape, and the resulting drive towards digital transformation that we are seeing affect almost every business in every industry.
We recently spoke with Kevin Goldsmith, CTO at Avvo and former Vice President of Engineering at Spotify. Prior to this, Kevin also worked at Adobe and Microsoft. Drawing on his experience from these businesses Kevin shared some of his past success and challenges, as well as his thoughts on how technology and engineering are impacting digital transformation efforts, and where he believes the future of digital lies.
Kevin Goldsmith (KG): I hear that phrase a lot, and I hear it used in a lot of different contexts and with many meanings. For a lot of businesses, I think it is now more about a change in priorities. It is about putting digital first in everything they do.
Because I have worked primarily in software companies I’ve never really being through a digital transformation as such because I’ve always worked in digital-first companies, so it’s interesting for me to see how companies that weren’t traditionally technology driven are now embracing technology in ways that are completely remaking their businesses.
KG: I think engineering and technology is penetrating all industries now, and really driving this digital first mind-set that we’re seeing. Whenever I meet with companies, in other industries, it is always interesting to hear how they don’t believe they can work in the same way software companies do.
For instance, think about how we can now produce software, ship it at will, and update it multiple times a day; when you talk to companies like legacy banking or health businesses this prospect is inconceivable for them. But, over the course of a conversation you slowly change their mind-set and show them how they can bring software into their business, increase their innovation and generally transform their business by being digital first and embracing technology and engineering.
KG: Being digital-first is fundamentally about being more responsive to your customers, either in the products that you build, the ability to customize the product, or bringing new ways for the customer to interact with your products or build a relationship with the company itself.
KG: There was definitely a continuum in both. I was at Adobe for 9 years, and when I joined we were very much still in the world of software being delivered on plastic disks that were shipped to stores. By the time I left, we were shipping mobile apps connected by services that were running in the cloud. Essentially, during my time there, the entire world of software changed.
One of the struggles we had at Adobe was moving away from a 12-18-month mind-set into a bi-monthly delivery, which we then eventually got down to monthly delivery. When I moved to Spotify it was about getting this down to three weeks and eventually, when I left, it was down to one week and we were always updating our services daily. So, again it was the shift in the speed of delivery and that ability to embrace customer feedback much faster, which was the real change during my time at those companies.
I got to witness the exhaustion of that long release cycle, and the excitement about being able to move faster and embrace new techniques and technologies.
KG: I’m definitely interested in the future of serverless, and the ability to orchestrate serverless in the sense of scripting your entire infrastructure and having serverless be part of that. I’m also interested in how the public clouds are taking technology that we previously spent months or years building and perfecting, and they’re basically making those APIs open. This is especially true around machine learning. We are seeing the commoditisation of machine learning technology.
There’s definitely stuff that is mid-term. For instance, I follow blockchain and its non-cryptocurrency uses for establishing provenance and things like that. In the longer term I’m looking to the likes of quantum computing to see if that will become realistic in the future.
Part of working in this industry is that there’s always new stuff to learn, so there are a lot of emerging trends I’m excited about!
KG: I think the rate at which technology is advancing, it is incredibly exciting and impressive, and this will only continue to open up opportunities for businesses. However, I do worry about the sheer amount of knowledge that people need in order to succeed; there is now so much to learn, and the breadth and depth of knowledge you need to be credible is increasing, and this only going to continue as technology evolves.
So, although rapid advancements in technology are a blessing, they are also a curse and I worry that it is only going to get more difficult for businesses to grow, as the talent they need just won’t exist.
KG: Definitely being able to scale the team – even more so than the challenge of scaling technology. As you go through rapid acceleration in a business like we did in Spotify you need to be able to maintain the good aspects of the company, while Spotify inherently had a strong engineering culture, when your team grows by 50% a year, maintaining it is always going to be difficult.
KG: Spotify was set up so that each team runs fairly independently of each other which was great, but one of the challenges of that model is when you have a large-scale project that needs collaboration.
One of the things I am most proud of during my time there was launching Spotify Now in 2015, which was an absolutely massive project and required hundreds of engineers. We were able to coordinate multiple autonomous teams in a way that wasn’t too disruptive and that resulted in the project being delivered on time and without asking people to work their weekends or nights. So, I’m proud that not only did we deliver the largest project Spotify had done, but that we were able to do it in a really good way.
KG: I think my challenges at Adobe were similar but for different reasons. I spent most of my time there working on the technology needed for carrying out efficient acceleration for image and video processing. That required a tremendous amount of technical innovation. It was a multi-year effort and I’m extremely proud of that from a technical perspective because it was probably the most complicated technology that I’ve ever delivered and leading my team through that was a privilege.
I was also part of the team working on Adobe Revel and we went from prototype to shipping product in 6 months – normally Adobe worked in 12-18-month release cycles so being able to change the culture to iterate quickly and work in this unusual and new way was an incredible experience.
KG: What will usually determine when it’s time for me to go is asking myself ‘If I stay here another year, what am I going to learn and what will be different?’ When I get to the point where I know that I’ll be doing the same thing in another year that’s when I start thinking it might be time to do something new.
KG: I think there is and there isn’t. One of the things we did at Avvo was working with some developers who came from non-traditional backgrounds and had gone through a boot camp or coding school. We had some very successful developers who were able to contribute a lot that way.
People are sometimes fixated on the idea that the four-year degree programme from the right university is the only way to get into industry. As a result, they tend to overlook people that could contribute significantly with a bit of training. It makes me think that there’s a skills gap due to how the industry evaluates talent, not because of a lack of talent itself.
On the other hand, when we look at the more advanced technologies that are the fundamental building blocks of software, I do see a bigger skills gap there. For instance, universities don’t tend to teach C++ anymore or some of the low-level languages simply because there is such a broad number of things to learn in Computer Science. But these languages are still the fundamental building blocks that underlie cloud and all of our newer technologies, so we still need this talent in our industry.
KG: I would actually say they don’t, but that’s because I see the skills that I needed when I joined the workforce – such as a growth mind set, curiosity and a desire to learn are still vital. I think the technical skills you need in the digital world are going to grow and change, but anybody entering the workforce needs those fundamental soft skills.
For me this has always been a core requirement, and even 10-20 years on you still need them. Maybe what’s changed is that every industry is a now technology industry because of digital transformation, so there is a new emphasis on these skills.
KG: If I look back to my 22-year-old self, I wish I that I had been a lot more deliberate and thoughtful about my decisions about my career. I focused on the name of the business, joining teams because the product was cool or because they used a certain technology.
It ended up working out find I think I made some mistakes along the way that could have been avoided just by being clearer about what I wanted, where I wanted to go, and the best way to get there.