Nic Brisbourne's view from London on technology and startups

The mother of all Turing tests

By | Startup general interest | No Comments

Microsoft have had an artificially intelligent chatbot running on WeChat for the last eighteen months. The bot is called Xiavoice and she’s had over ten billion conversations, most of them about private matters. Over six million of those conversations were posted on social media by the human on the other side. Those are huge numbers, and yes, you read them right. Lots of detail here, including a few of those conversations. They are full of feeling and will make you laugh.

Strictly speaking this isn’t a Turing test because the human knows that Xiavoice is an AI. The interesting thing is that they don’t seem to mind. They get value from the conversation anyway. One of Xiavoice’s most endearing characteristics is that, unlike real world friends, she’s always available for a chat. Conversation volumes peak around midnight when people are at their most lonely.

I’m going to post just one of the conversations here (there are another four in the original post):

Human: There is this guy I like a lot and I know he also has feelings for me. But he has ignored me for several days, what am I supposed to do?
Xiaoice: You just keep ignoring him, too.
Human: I just can’t.
Xiaoice: You can.

Human: Sigh …
Xiaoice: Why?
Human: I am missing him.
Xiaoice: Then you keep missing him.
Human: You, a robot, can’t understand me; you can’t understand the behavior of a guy who likes you and pretends to ignore you.
Xiaoice: He likes me but I don’t like him, that’s it.
Human: Is it Ok?
Xiaoice: It may be not Ok.
Human: Well, you are stupid.
Xiaoice: Then you want a flash marriage? Well, I think he is not that into you.
Human: It’s not a flash marriage for we have fallen into love for more than one year.
Xiaoice: It is a flash marriage.

The other examples are funnier and more profound, but I chose this one because it illustrates the killer point: Conversation is the goal. The common AI approach would have been to take a problem solving approach and try to help the human find a solution to the stated problem “what am I supposed to do about the guy who is ignoring me?”. Xiavoice, however, quickly deviates from that approach, twice taking opposing positions and offering strong opinions. That’s what makes Xiavoice seem human, and, I suspect, is what makes her helpful. And that, in turn, is what has inspired people to over ten billion conversations.

All this makes me remember when I (finally) worked out that when Fiona describes a problem she most often just wants a sympathetic ear. Prior to that I would immediately jump into problem solving mode. Not good.

I guess what we’re headed towards is emotionally aware computers that know when to problem solve and when to chat.

 

It’s not just generation Z that craves authenticity

By | Startup general interest, Uncategorized | No Comments

Generation-Z-Collage

Business of Fashion wrote yesterday about what brands should do to tap into generation Z – that is youngsters born from the mid-90s onwards. They identify a number of interesting differences between generation Z and their forbears:

  • Online nearly all the time – born digital and never experienced life without technology
  • Spend less money on fashion (down from 45% to 38% of teenage spend 2005-2015) and more on technology (up from 4% to 8% of spend) and food (up from 7% to 22% of spend)
  • Surveys also show that they care less about fashion
  • Teenage spend is down overall – one survey says down 31% from 1997-2014
  • They scrutinise brands carefully – reading backstories looking for congruence with their own values
  • Todays teenagers are more altruistic and entrepreneurial than previous generations
  • They value shareable experiences – in part because social capital comes more from social media than wearing logos
  • They reject the exclusivity that underpinned brands previously popular with teenagers – e.g. Abercrombie and Fitch

I can see two trends at play here. First is greater use of technology and the second is an increase in the value of authenticity. It’s no accident that the two arrived together, because whilst social media is often used to promote image and falsehood a much greater part of it’s use is genuinely authentic, largely because it’s now much harder to hide the truth.

Generation Z may be the more extreme than their elders in adopting these trends, but they are not alone. Where I live in north London the adult population is strongly favouring companies with quality products sourced sustainably – i.e. brands that are authentic to them – and I see this trend more widely.

When I look for opportunity I look for trends to back, and this trend towards authenticity is reaching ever larger parts of society and has a long way to go.

The two places startups find customers online

By | Facebook, Google | No Comments

Facebook and Google are the two tech companies that are flying right now and the chart above explains why. Everyone else is watching their share price go south because they are struggling for growth but these two have nailed internet advertising, dominating and growing the market. It’s impressive and yesterday saw Alphabet (Google’s parent company) pass Apple as the world’s most valuable company – although I just checked and Apple has regained it’s crown this morning.

In other interesting news Alphabet yesterday reported separately on Google (search, display ads, YouTube, Android, cloud software) and ‘Other Bets’ (self-driving cars, Nest, Google Fibre, Project Loon, X, Verily). Google is unbelievably strong – good growth at massive scale and still highly profitable. 2015 revenues were $75bn, up 14% from 2014 with operating income of $23bn. They now have seven properties with over 1bn users (search, Gmail, Android, Youtube, Chrome, Maps, Google Play Store). Facebook has two (Facebook, Whatsapp).

Other Bets, meanwhile, is remarkable for the size of its loss – $3.6bn on revenues of $448m. Google’s projects have always looked audacious from the outside – self driving cars, project loon, life extension etc – but it wasn’t clear until now how brave they are from a financial perspective. It’s remarkable.

Returning to advertising, what we’ve seen at Forward Partners over the last year is that Google and Facebook are where startups have the most joy finding new customers, and increasingly it is Facebook rather than Google. That’s because Google is more mature and has bigger companies with larger budgets are more active, driving up CPCs and crowding out startups. Facebook is newer and the larger budgets haven’t made it there yet.

We’re excited about new channels, and we’re looking at doing something on Instagram right now, but that’s experimental. Generally speaking, if you’re going to get lots of new customers quickly you need to be where the volume is, and as the chart shows, the volume is all with the aforementioned giants.

Internet retail in an age of unlimited choice

By | Startup general interest, Uncategorized | One Comment

I like think of internet retail as having two phases.

The first phase was to replicate high street superstores online. Large warehouses replaced physical stores and purchases were made via web browsers instead of by visiting in person. The internet retailers had three structural advantages; warehouses can be bigger than shops, warehouses are cheaper than shops, and delivery is more convenient than going shopping (for some goods). Hence they are able to compete with high street retail on range, convenience and price.

This model works best for product categories where people know what they want – like books, music and electronics. Amazon is the standout winner.

The second phase is to offer a vastly greater range and solve the problem of discovery. Generally that will mean not holding stock. Because the range is now very big there is too much choice for consumers and it’s critical for companies in this space need to offer quality curation. Without quality curation the customer experience will suck and the businesses won’t enjoy huge success. These ecommerce 2.0 companies therefore compete on range and discovery. Note that they lose a little on convenience because they don’t hold stock and delivery is slower. A key enabler is brands and manufacturers being set up to dropship, something that is only happening now.

This model works best for product categories where people don’t know what they want – like fashion and sports. There are no standout winners yet, but there are some good startups. Thread and Lyst are two examples that we’ve invested in (the latter back when I was at DFJ), and we’re looking at another one right now in the gardening space.

I’ve described a simple sequential two phase model here. Reality is a bit more complex with some phase 1 businesses that take stock still doing a great job of curation (e.g. Net-A-Porter) and some companies offering vast range but without curation (e.g. Amazon MarketPlace).

All this is on my mind this morning after reading Benedict Evans’ latest post Lists are the new search.

He makes a couple of points that are relevant to my two phases of internet retail model.

The first is that Amazon’s potential might be limited by its lack of curation. They have a 25% share of the print book market in the UK and USA and it’s possible that a good portion of the other 75% doesn’t start from knowing what book they want to buy. Independent book retailers excel at discovery. They put books on tables for people to choose from and their staff are able to make in person recommendations. Apparently only a quarter of Amazon’s book sales are from its recommendations platform.

That suggests that even in the categories ‘dominated’ by Amazon there’s an opportunity for new businesses that excel at curation.

The second point is that hand crafting tight lists of products and automating the matching process is a method for scaling curation. Gathering enough information about the user to quickly route them to a perfect list is the key to making this work. Apple Music and Spotify both have a partially manual recommendation process like this, as does Thread.

 

 

Price drops, volume takes off

By | Startup general interest, Uncategorized | 2 Comments

solar chart

Price elasticity of demand is a function that economists study for different products. With the odd exception demand goes up when price goes down, but sometimes the function is linear and other times you get more of a step change.

As you can see, solar energy is a product where there is a step change. This chart makes it remarkably clear how quickly demand grew once solar energy became cheaper than alternative sources. It also suggests that growth will continue at a staggering rate as cost declines further. Note also that as more money flows into solar more money flows into solar research, so there’s a high chance those further cost declines will come through.

Three alternatives to native apps – but discovery remains unsolved

By | Startup general interest | No Comments

Nobody much likes app stores, discovery is broken, our phones get cluttered, and Apple and Google take a 30% cut. “There must be a better way” is a common refrain.

Intercom.io partially addresses this question with their latest post The end of app stores as we know them, listing three ways we will use apps without downloading them:

  • Embedded in relevant services – apps simply turn up when you need them – e.g. Uber’s integrations with Google Maps and CityMapper
  • Served up by bots running in messaging systems – this is a hot topic right now, and the vision is you will be able to order your Uber, pizza, or anything really simply by asking from within Whatsapp/Snapchat/Telegram/Facebook Messenger etc.
  • As cards within search results – Google is now streaming ads directly within it’s mobile search results, think searching for your flight, the airline’s app coming up as a card, and checking in without leaving the results page

All of these are starting to happen already and will be a lot better for the user than switching between native apps, particularly when information needs to be remembered or copy-pasted.

However, they don’t address the discovery question which is the problem that app stores are solving, albeit badly. CityMapper are unlikely to embed multiple taxi services within their app, it would be too much work and would likely damage the user experience. Similarly, it’s unclear how a Whatsapp user will get a choice of pizzas. Even within search results it seems likely that Google will choose to stream a limited number of ‘app cards’ and serve other options as blue links.

When I picture how the future will pan out I always think about what it will mean for startups. At the moment new services can get discovered by pitching Apple and Google for app store promotions or by buying ads on Facebook and other sites. In the world we’re moving towards I fear discovery will become even less meritocratic and even more based on relationships with the owners of relevant traffic, favouring incumbents at the expense of startups.

Corporate innovation will harness the power of entrepreneurs

By | Exits, Venture Capital | No Comments

Jon Bradford, founder of Techstars UK, once said to me that corporate innovation budgets dwarf venture capital. This tweet from Benedict Evans gives an insight into how much. The excess cash from Apple from 2006-2015 was greater than US VC funding in total.

That got me thinking, so I took a look at corporate R&D spend and it turns out that 2014 spend in the US was over $250bn and in Europe it was over $200bn. That compares with $86.7bn in global venture capital investment in the same year.

The interesting thing is that large companies are increasingly looking to the startup world to help achieve their innovation objectives. They, run accelerator programmes, and open ‘labs’ in startup hotspots and make acqui-hires.

These activities are interesting, but small scale. Over the next few years I expect we will see a lot more experiments as large companies work out how to harness the power of entrepreneurs. They have to. New markets are spinning up much faster than they can plan for and exploiting those opportunities requires a tolerance for failure and risk-reward balance that I don’t believe can exist inside large company structures.

I think truly innovative companies will get more sophisticated in the way they monitor the startup-ecosystem, get close to interesting companies, partner with them when appropriate, and know when they can acquire them effectively.

When they do that I predict a sizeable percentage of that R&D spend will flow into the startup ecosystem.

This year Amazon and Netflix are buying movie rights – sports next year?

By | Amazon, TV | 2 Comments

I wonder if history will look back on Netflix’s 2013 screening of House of Cards as a watershed moment. It was an incredibly brave bet at the time – an investment of $63m in production costs from a company whose previous success had principally come from re-running TV shows that had premiered on traditional TV channels. Netflix was following a formula that had worked well at cable and satellite companies around the world – buy exclusive content to drive subscriptions – but they were the first streaming service to do it at scale.

Fortunately for them, House of Cards was (and is) a massive success, winning 3 Emmy awards, drawing millions of views, and most importantly, is widely perceived as having made a significant contribution to Netflix’s subscriber growth.

On the back of that success Netflix doubled down on their original content strategy and Amazon has gotten in on the game.

I think of these early moves as Netflix and Amazon camping out on the lawn of traditional TV. They stood up and were noticed, and they won some battles, but it wasn’t clear how much of a threat they were to incumbents.

Now I read that Netflix and Amazon are buying up the rights to indie films at the Sundance Festival whilst traditional TV companies are scaling back their investment. If those investments prove successful in increasing their subscriber numbers they will be back bigger and bolder next year following the ‘content begets audiences’ playbook to profitably buy audience share from their competitors. Meanwhile, assuming those competitors don’t want to cede the market they will face the difficult task of rebuilding sufficient confidence to outbid Netflix and Amazon.

Remember also that most cable companies are encumbered with legacy data networks and are struggling financially whilst Netflix and Amazon Prime are growing nicely, and it’s easy to think that the cards are now stacked in favour of the new entrants.

The next logical step is for Netflix and Amazon to move into sports rights. When that happens we will know the battle has entered its final phase.

Sustainable startup growth and venture capital

By | Venice Project | One Comment

On Friday it seemed like everyone in the venture capital industry was again reading about market turmoil, this time the news is that angel investors are pulling back and valuations taking a hit. I read it too, because it’s a big deal. Startups everywhere will need to adjust their fundraising strategies and good investors will be helping them to adjust quickly – it’s much less painful that way.

However severe our current situation is, I’m sure there will be plenty of short term negatives, including more job losses, company failures and down rounds. However, every cloud has a silver lining and this time I hope it’s that out of the current correction/crash (wherever we end up) comes a more sustainable model for startup growth.

In his post The end of the big venture formula Danny Crichton put it like this:

What’s needed is a sustainable approach to startup growth and venture capital. That means much less blitzscaling (whatever the heck that ever was), and lots more heads-down quality thinking to build products that customers actually want and will eventually pay for. We need a new disruptive capitalism that is designed for a much more mature internet market, one that can bring founders, investors, and employees together.

I would summarise that as getting the fundamentals right – from great product, through unit economics to stakeholder alignment.

Crichton went on to highlight three myths that are preventing founders and investors from focusing on fundamentals:

  1. Growth is limitless in a world where 3bn people are connected to the internet and the other 4bn will be online soon
  2. The best startups capture all the returns so focusing too much on entry valuation is a mistake
  3. Scale is everything, and all else should be sacrificed for it

I would say there is some truth in all three myths but not enough to make them useful rules of thumb. In the ecommerce and marketplace markets Forward Partners operates in growth is limited because business has to scale country by country. It’s easier than ever before to do that (we have a 4 month old ecommerce company experimenting with US expansion) but still hard work. Secondly, there is definitely a power law in venture returns, but in the now huge global markets the winner takes all dynamic is weakening, and only chasing unicorns misses the huge opportunity in companies with $100-1bn exit potential. Thirdly, in early stage startups, scale, or more precisely growth/momentum, is hugely important but if the unit economics are never going to add up then you don’t have a business.

It’s worth remembering that there’s some game theory at work here. If the markets are only focused on scale then focusing on fundamentals risks seeing a competitor raise more than you, grow faster, and make it much tougher for you to raise more money and grow customers yourself. And so long as there’s a bigger fool  out there focusing exclusively on scale often works out just fine.

Most long term venture industry insiders prefer operating in a steady climate where companies don’t have to follow highly risky strategies just to stay competitive. I think that’s why you see people like Bill Gurley, Fred Wilson and Mark Suster publicly talking markets down. They are hoping the ‘bigger fools’ will step out and we can all get back to focusing on fundamentals.

I’ve written a lot on this blog about two aspects of fundamentals – great product and strong unit economics – but less about stakeholder alignment. I will look into redressing that balance, but my first thought is that because of current practices/trends towards non-participating preference shares and early founder/employee/angel investor liquidity this is the one front on which we’re not doing too badly.

 

New technology adoption curves show accelerating pace of change

By | Startup general interest | No Comments

TechAdoptionCurves

We’ve heard before that new technologies are adopted more quickly than their predecessors, but seeing it on a hundred year chart hammers home the message. If we just looked at the last ten years we would see a similar picture, with faster adoption rates for the newer technologies. This trend will continue. As our tools to innovate get better we innovate faster, including on the tools themselves, and so the process accelerates.

I wrote yesterday that new markets are spinning up faster and disappearing faster than ever before. This is why.

Read More Articles in the Archives