Monday, 21 March 2016

Thoughts on SXSW 2016

Don't like words, just like pictures? Then check out this technology-focused photo album compiled by myself and Lorenzo Wood, or my broader Austin photos in this Flickr album. Warning: it will make you hungry.

Austin was magnificent, as always. A mix of great company (cap doffed at Warren, Dave, Gavin, Lorenzo and others), amazing food (BBQ at Kreuz Market, amazing steaks, OMG brisket) and terrific entertainment (The White Horse, Pedalo Karaoke, Buffalo Billiards) adds up to a winning combination. Was fab having so many DigitasLBi colleagues there from around the world too.

As for the actual sessions, there are over 1,500 to choose from now, no really, which not only means you can only attend a tiny percentage, but also means you end up choosing the ones you want based on session title only. Who has the time to read 1,500 synopses, never mind thousands of speaker bios? I heard quite a few grumbles about the quality of the talks, but it's impossible to manage. Lanyrd used to be brilliant at crowdsorting through this mess, helping sort the wheat from the chaff, but that's only worthwhile if everyone's using it. Hopefully the Upcoming renaissance will be ready for next time.

The net result was that the sessions I attended were an equal spread of awesome, good and poor. But the awesome ones alone made everything worthwhile, especially those exploring the cutting edge of AI. Obviously AlphaGo's success was the talk of the town, but it was the nature of the success that was so exciting. By making moves that humans couldn't predict or understand, it gave an insight into our machine-dominated future. 

In the short term, we can expect ongoing examples of narrow excellence, much as we've seen already going back to Deep Blue and Watson, but coming thicker and faster. The big difference between these and AlphaGo is that we didn't teach the computer how to play Go, we taught it how to learn. As we, and they, get better at this, the growth becomes exponential, especially when we invite these learning machines to deal with systems that are too complex for humans to process or understand. 

As with AlphaGo, they'll probably find insights and opportunities that are beyond our comprehension, leaving us with the question: when should we start to put our faith in them?

Plus the way these systems overlap with biotechnology, genetics and nanotechnology is truly scary - and not in a good way. I still highly recommend the Wait But Why article on this. 

I digress. Back to the talks. I particularly enjoyed the talk given by Dag Kittlaus, whose team developed Siri. He's now working on a system which not only lets you upload data sets into the cloud for their systems to combine and learn from, but which also has a voice interface. He imagines a world where his 'v' logo is as ubiquitous as the Bluetooth logo - when you see it, you can talk to it. An incredibly smart and engaging speaker. Check out to find out more.

What else? Obama was excellent (no I didn't win the ballot and see him in person) - watch the video when you get a minute. JJ Abrams was good, talking about imbuing robots with human characteristics to drive empathy. Kevin Kelly was engaging as always - he talked about pills that can measure you from inside your body, and dictate the contents of the pill for the following day. Far out. 

The keynote by the Under Armour CEO (Kevin Plank) was pretty inspirational too, albeit in a chest thumping kind of way. His main message was that it isn't good enough to have a connected ecosystem that tells you about past or present events - it needs to guide you on the future as well. We already have constituent parts (apps) that help us understand diet, exercise, sleep, wellbeing, happiness and more - but as yet nothing stitches it together very well to help you make smart decisions in real time. I believe the companies that get this right across each sector have a bright future.

And VR was everywhere. We're clearly at the tipping point. Check the photos for some interesting applications, including how you can trick people into walking around a virtual environment that's larger than the room you're standing in. 

One final personal note. It's three years since I was last at SXSW, but the combination of Uber plus Vodafone's $5-a-day roaming plan made the whole experience so much better. No more waiting for the shuttle bus, no more missed sessions, no more waiting for an hour at 3am for taxis, no more worrying about missing flights. RIP Austin taxis. 

So that's it. SXSW isn't without its detractors, but it brings to mind the old Yogi Berra quote - nobody goes there anymore, it's too crowded. As a conference it's challenging to get the best out of the experience, but if you relax a little and strike the right balance between sessions, entertainment, the trade hall, networking, socialising, eating and generally soaking up Austin, it's still the best nerdfest the world has to offer. Viva la South-By!

The path to success

This post originally appeared in Campaign Asia on 8 March 2016.

Everyone wants to channel the incredible disruption and success of Silicon Valley, but how should we go about doing it? Whenever great leaders write their memoirs or business self-help books, common themes tend to emerge. The path to success is rarely straight, usually featuring a fair dose of good fortune, adversity, hard work and close calls that make the story compelling. And the companies involved often share the personalities of their leaders.  These companies, particularly those in the technology sector, often bear the hallmarks of resilience and adaptation that should see them flourish for years to come.
But when we look inside the typical organisation things aren’t so encouraging. We can see that many processes and practices are designed in such a way that they limit exposure to the conditions that drive success. A new business forged in harsh market conditions can only succeed through a maniacal focus on customer needs, so why do so many companies restrict this frequent contact with their customers?
It isn’t all bad news. Some processes, such as those employed in product design, have established methodologies to drive user engagement and validation, from inception through to launch. Maybe this is because failure rates are generally high if customers aren’t directly involved along the way. But others, particularly in the marketing space, fall short. From strategy to design to production, unvalidated assumptions are made about the customers’ motivations and behaviours, usually through rose-tinted glasses. The team works for months in an echo chamber, and then wonders why the end result falls flat. The failure to engage with customers across the board leads to a business which is less effective and ultimately more exposed to an unforgiving market.
Why is this so? Are the ideas so majestic and Earth-shattering that any member of the public would immediately break any non-disclosure agreement to get the news out? Of course not. In fact from experience I’d say not involving the customers makes the opposite likely. Assuming the leader has been brave enough to back an innovative idea, it tends to get watered down between agreement and launch as the lawyers and jobsworths have their say — unless the customer’s voice keeps the idea alive.
Given how many successful people and companies talk about testing and learning and iteration and validation, it’s remarkable how few companies actually put it into practice. Remarkable — but understandable. It’s time-consuming, can be expensive, impacts your potential to hit deadlines and it forces you to confront criticism. And those are just a few of the reasons.
But when we start to pull the thread of those excuses, they start to unravel. Testing with customers can be expensive, but it doesn’t have to be. By putting effective constraints around the ideas being tested, for example using paper sketches rather than production quality assets, and insisting on quick and fast tests, we can prepare in a fraction of the time and cost. Products and services such as, Testbirds, Marvel and Invision can help.  And as it turns out, the less time you invest in the assets, the better you take the criticism because you haven’t invested too much of your soul in them.
This one constraint can unlock a huge amount of potential. By testing early, we’re getting incredibly rich and valuable feedback from the marketplace. This doesn’t just help us to improve our idea, it can also help open up completely new avenues for exploration. It’s this repeated exposure to the market that helped Odeo pivot into Twitter, helped Game Neverending pivot into Flickr, turned Burbn into Instagram, and many more.
You might still find you can’t hit your deadlines, and that’s OK. Much better to know you’ll be failing early on in the process, and pursue an approach with far better prospects for success.
In short, having a boss that supports and drives innovative behaviour is the key to unlocking this huge opportunity.
One concern that has to be addressed head-on is the huge percentage of startups that fail. It’s said that somewhere between 75% and 90% of startups fail in their first three years, and it’s fair to say no-one’s career is going to last long with that kind of strike rate. But this is where comparisons with Silicon Valley and the world of start-ups isn’t really that helpful. You don’t need close inspection to see that the starting conditions are having a huge impact on the outcome. Large organisations have an existing successful business model and customer base that would be the envy of any start-up. And by their very nature, start-ups are typically bringing an untested product or service to market that has to successfully jostle for space in the crowded lives and wallets of their potential customers.
Protecting brand equity is also an interesting area to explore. You could argue that their nothing-to-lose attitude gives start-ups free reign to explore ideas that could cause irreversible damage to an established company. The bigger the brand, the greater the fear that a small mistake can have catastrophic repercussions - and the higher the chance that risks won’t be taken.

We can see there are dynamics at play within the start-up world that are appealing to established companies, but adopting them wholesale is unlikely to get the results we’re looking for. By carefully adapting them for the corporate environment, and by getting closer to our customers, we can plot our path to success.