Main illustration: Blake Kathryn
The relentless march of technological improvement means that by their very nature technology businesses fail.
That’s why as founders, product people, marketer – or whatever our role is – we need to be acutely aware of all the different technological shifts happening in the industry and consistently ask ourselves if and how these things actually affect us? This is not just some dry strategic exercise. If you don’t do it your business could die an awful lot quicker than you expect.
It is the nature of our industry, it is the nature of technology, that every product dies. That’s a very morbid thought, but if you don’t believe me on that, page me, or fax me, or write me up something on your typewriter. They were all epic businesses once upon a time, but they are no more.
Death is on my mind for a lot of reasons. It is the nature of every technology to die. It is the nature, therefore, of a lot of technology businesses to die, which leaves you with an eternal question – how can we stay alive?
I wasn’t always so morbid. When myself and my cofounders Eoghan, David and Ciaran signed Intercom’s incorporation papers in San Francisco in 2011, we were pretty optimistic back then. Intercom has been something of a roller coaster. But a very good roller coaster – real Six Flags stuff.
Above is a graph of one of our metrics – homepage traffic. There’s been a few different phases at Intercom, and the ones I can roughly identify are the phase when the product is being born, the phase where we had to then go and grow the product, and lastly the phase where we have to survive. The latter is where we’re thinking: “Okay, we’ve made something. How do we stay relevant in technological cycle after cycle after cycle?”
Put another way, we have to make it work, we have to make it grow, and then we have to keep it relevant. We have to give folks a reason why they should continue to use Intercom. Today we have more than 15,000 customers, tens of millions of dollars in revenue, and $116 million raised. All of these things should be cause for celebration, but the reality is we’ve never felt more vulnerable. We spent so long hoping something could come along and exist, that when it began to exist we were like, “Holy shit, what if it goes away?”
I get that sense in part because I’ve put a seventh of my life into Intercom (which makes me 35, I’ll save you the math). That’s a significant chunk of time to work on anything, and of course the idea that it might one day no longer be around is terrifying.
The weird thing is, this isn’t what we were sold – when we read all the books about startups, and heard how it’s supposed to work. I was always thought the deal with startups was like The Shawshank Redemption. You crawl through 100 yards of shit, and you ’re free. But what it actually is more like is you crawl through 100 yards of shit, and then your options are, much like Andy Dufresne said, “Get busy living or get busy dying.” How do you stay relevant? How do you stay alive?
Why only the paranoid survive
When you look at all the literature, all the advice, all the blog posts, the articles, the podcasts, the conference talks, it’s all about creating and starting. Because everyone wants to do that, that’s what people write about. At Intercom we’ve added our own content and talks to that along the way. But in reality none of this matters if you’re ignoring existential threats. If there’s some reason, some ticking bomb or a train coming down the tracks, and you don’t care because you’re too busy navel-gazing, you’re going to be in trouble. On that note, I really like this quote from Andy Grove, when he says, “Any degree of success will breed complacency. Any degree of complacency will breed failure. Therefore only the paranoid survive.”
I recently shared these thoughts with a friend who is also a startup founder. The conversation went something like this:
Friend: “Well the thing you have to realize is that the present is the past.”
Me: “I don’t do drugs, so you know …”
Friend: “No no no, listen to me, listen to me. Your product’s already obsolete.”
Me: “What do you mean by that? People like it.”
Friend: “Well, the thing is, people like what’s out there today, but what’s out there today is actually already history. It’s been documented, there are “changes” pages updated, but it’s live, it’s now a part of history. The thing you should be worrying about is that there’s probably going to be a newer, faster, quicker way for your customers to achieve what they want to do in this world, and that it doesn’t involve you.”
Me: “Ooh, that kind of stung.”
The way he described it to me, it’ll feel like you’re on solid ground and everything’s going great. You’ll feel a couple of trembles here, and a competitor there, and a couple of different shifts. But you’ll be saying, “We got this, we got this, we’re good at product.” At some point you realize something’s actually changing, so you start trying to react. But by the time you actually realize it you’re too late. The ground has been torn out from under you, and the worst part is you don’t feel it. It’s purely asymptomatic. If you don’t believe that it can happen, even when times are good, let me put it to you this way: Wasn’t 2007 a great year to be a satnav manufacturer?
Things could not have been better. We were all buying them and putting them in our cars, and they were shit devices, but we didn’t care. Then, somewhere in the Moscone Center in San Francisco, someone waves a new thing around, and poof…business no more.
Both Garmin and Tom Tom have not restored their former value since the introduction of the iPhone – at best they’re a sixth of their peak. It’s a fun version of history to presume they were all literally asleep at the cash register, totally ignoring all these new threats. The reality is I’m sure they watched the keynote and panicked. I’m sure they ran straight to their whiteboards. They probably pulled the product team in on weekends. I’m sure they fired up JIRA, and when it loaded up they’ said, “All right, let’s go and start writing some tickets.” But customers don’t wait around while you’re writing up JIRA tickets. The world will move on, with or without you.
Another simple example: In the graph below, the light-green line represents a product called SMS. Aside from exorbitant roaming charges, SMS was a phenomenal technology. SMS is the only way telcos make money. It’s like liquid profit.
Then 30 engineers got together and built a different thing called WhatsApp. Within the space of two years they’d basically obliterated the majority of the growth in the most profitable part of telcos. Now, it’s important to say I am talking about growth here, because businesses can’t survive on stagnation. They need to grow, because ultimately, on a long enough time frame, all their customers will die, so you do need new customers. You can’t flatline as a business. Flatline actually means you’re going down.
The way this plays out typically is a new threat emerges and the incumbent typically says, “Oh yawn, someone started a product, who cares?” Eventually WhatsApp puts out a press release saying, “We’ve just got our 100,000th user,” and the telcos are like, “Hahaha, 100,000 users, what a joke.”
Then at some point WhatsApp puts out a press release saying, “We have our 100 millionth user,” and the telcos think, “Oh.” Then by the time they decide to fight it is literally already a done deal, and there’s no turning back. If that framing sounds familiar to you, it’s because it’s the words of Gandhi, who’s obviously got a lot of great quotes and isn’t necessarily a product manager. One of the points he made was this idea that, “At first you’ll ignore them. Then you’ll laugh at them. Then you’ll fight them, and then they will win.” That tends to be people’s attitudes towards new products.
Know your place in the technology constellation
If you want to see the best example of what it means to first laugh at something, a new threat, I really like Steve Ballmer’s reaction to the iPhone.
If all that the mobile revolution, the new Android and iOS things, did was kill Windows Phone, that’d be fine. Windows Phone wasn’t a huge thing for Microsoft, it was a part of their Windows Everywhere strategy. But mobile didn’t actually kill Windows Phone. It attacked the whole entire desktop concept, which scares the hell out of Microsoft because that was their core.
It wasn’t obvious that a phone could be really, really bad news for a desktop operating system, but it turned out to be. The realization here is best summed up for me by this quote from Steve Sinofsky, who is ex-Microsoft and is now at Andreessen Horowitz: “No technology is the center of a system, but rather a constellation of bodies under the influence of each other.”
“No technology is the center of a system, but rather a constellation of bodies under the influence of each other.”
He makes the point that all these technologies just intertwine with each other. It’s not clear that a phone was going to destroy a satellite navigation company. It’s not clear that messaging software would destroy a telecoms company. It’s not clear that a phone would disrupt desktop. But what happens in technology are these tectonic shifts, literally the plates slide from under you, and if you’re not aware of all of them and how they interact, you’re in big trouble.
To go back to Intercom for a second, we didn’t come this far to only come this far. How do we actually keep going? How do we make sure that we don’t get caught up in one wave and then finish up?
Technology will continue to spit out innovation after innovation after innovation. The question you have to consistently train everyone to ask: does this new technology that’s happening or that’s being released make it in any way cheaper, faster or easier for our customers to make progress in their lives? That’s the repetitive question you have to ask, whether you see Bluetooth or WiFi or cloud or mobile or touch or voice or audio or messaging or bots. You name it. Because if it does make it cheaper, faster or easier for customers to make progress, they’ll go there, and you’ll be busy writing up JIRA tickets.
Needs don’t change, even when technology does
Connected with this is that the things we need to do in our lives rarely actually change. But the ways we can do them will always change. There are very few new behaviors in life. To give you some random examples, for as long as there have been kids in schools there have been kids passing notes to each other. Protocol with notes is that you make sure no one else can see it, you rip it up or you eat if the teacher catches it, but you just get rid of the thing. The idea that this was a disposable message that can be passed between two people for their eyes only, has gone on today, and we know it best as SnapChat. One of its core purposes is that exact job.
Those of you who are closer to my age probably had a box that you put photos in under your bed. These photos, curiously, are not on a shelf for everyone to see. They’re specifically for you, and you only want them at certain times. Today we use technologies like iPhoto or Dropbox’s Carousel, when it was around, to do this exact same job. The need has not changed, but the tool has.
If you’re renovating a house, in the old days you’d have a scrapbook with all the stuff you can’t afford. Today you do it with Houzz, or Pinterest. The need doesn’t change. The ways you can do it always change.
The way to stay relevant is to pay attention to what is called the OODA loop. Can you Observe, can you Orient, can you Decide, and can you Act? If your OODA loop is fast enough that you can keep up with the industry, you will always be in a great position. However if you are slower than the industry then you’ll be in trouble.
If you’re faster, every time the industry spits out something new – cloud, for instance – you can release something that works with cloud. You can release many iterations, and you can consistently evolve and build new stuff. The very second mobile comes out, you can build mobile straight away, because you can move as fast as the industry, or faster. That’s really, really important.
But the very second this arrow changes direction, and you find yourself moving slower than the industry, it’s game over. They’re spitting out technology after technology after technology, and your team is still talking in years and Q4 2018 and stuff like that. As the technology changes, you then find yourself being the company who in 2016 says, “We’re now available online.”
What that looks like is what Marshall McLuhan described as walking backwards into the future. You are moving into the future whether you like it or not, but your vision hasn’t changed. You’re still trying to force old-world ideas into a new world. That’s why it constantly comes back to this question: every time something changes, can we be cheaper, faster, or easier?
Part Two: Why you can’t ignore technology trends
Does your product make it cheaper, faster or easier for our customers to make progress in their lives?
When I look at 2016 and 2017 and think about our roadmaps, and I’m sure you’re all thinking about your roadmaps, there’s a few trends I’ve been observing. These will have ramifications for all of us whether we like it or not, because we don’t get to pick the future.
Watch me deliver part 2 of this post as a talk on the 2016 Inside Intercom World Tour.
Consolidation of the enterprise
The first one is the consolidation of the enterprise. We used to have these bloated multi-CD desktop software products and then frameworks like Ruby on Rails and Django came out, and everyone made a very curious choice. They sliced off a very small piece, a minimum viable product, of specific parts of these tools and released them as individual products. Where once you had an accountancy suite, now you have time tracking, and invoicing, and payments processing, and four or five other products. That would be fine except all these things don’t necessarily interoperate. Now for someone to get their job done, instead of using one product, they might use seven or eight because all of these things, they’re all interlinked.
This is a very Intercom-like notion because when we were talking about how can we explain what Intercom does, our CEO drew a whiteboard diagram with something like the above. It explained that on the left was the world before us. There’s all this different data, and the dream of Web 2.0 was that we’d spend the best part of our years writing API to API connections so that all this stuff could talk to each other. Of course, that doesn’t happen because you’ve got better things to do. We figured we could position Intercom as being the “anti” of this.
We sent this over to our brand design team, who are way better at diagrams than I am at Keynote, and they produced this. This became the iconic Intercom image for quite a while and we still have versions of it today.
What’s interesting is this problem wasn’t a unique thing to Intercom. This diagram was, at the time we released it, but a curious thing happened: Another company released a similar diagram, except for them it was about resumes and networking and online presence. Okay, fair enough. We saw a few more, and obviously, they got a little bit wackier along the way. We’re thinking, all right, maybe this is just a general trend and we just happen to have the iconic diagram for illustrating it. These days, however, I actually can’t go online without seeing this damn thing. It just keeps happening.
Clearly this is a problem people are experiencing – this idea that we over-fragmented when we went to the cloud and a lot of workflows have become disjointed across too many different areas.
What does this mean for you? Well, the questions you have to ask are:
- Does consolidating our tools make it cheaper, faster, or easier for our customers to make progress in their lives?
- Specifically, is the ease of consolidation, i.e., having this sort of centralized place, better than the benefits you’ll get from specificity by using single purpose tools for every single thing? If it is, you should really think about the implications of that.
- Are there adjacent workflows you should include? Do people, the very second they leave your product, jump somewhere else just to continue what they were doing? If so, someone’s going to eventually connect those dots, and if it’s not you, you will be in trouble.
- Similarly, do your customers frequently jump from product to product just to get one task done? For example, does paying an invoice involve having five tabs open and an iPad over here, or is there a singular way to get it done?
The enterprise is consolidating again, and there will be repercussions. I firmly believe we’ll see a lot of this. All those companies, sure, a lot of them will die, such is the nature of startups, but a lot of them are going to succeed, too.
Another trend that’s popular today is AI. It goes through this sort of love/hate cycle with the media, and in a lot of ways is everyone’s favorite whipping boy. I tweeted this on Valentine’s Day, because, obviously, that’s what you do on Valentine’s Day.
I’m increasingly convinced that most start-ups boasting about AI are really just impressed with their 183 if-then-else statements.
— Des Traynor (@destraynor) February 14, 2016
That’s what it feels like a lot of the time. AI hasn’t really, even today, delivered what we were promised. Now, it has delivered some cool things. For example, I saw a robot that can apply lipstick perfectly. Here it is.
It can also wake you up early in the morning, which is pretty handy.
My favorite piece is after it wakes you up, it can actually make breakfast for you, which is probably the best thing it has to offer.
That feels like the future that we’re fighting for, doesn’t it? It is genuinely getting somewhere. A car company recently released the video below. It shows a car pulling up, projecting a pedestrian crossing, and letting the pedestrian cross.
They finish this video and it fades to black and the question is simply, is this a concept video? You’re thinking, of course it’s a concept video. The answer is, actually, it isn’t. You’re like, oh, shit. Maybe this stuff will come.
I genuinely believe, as in the case of self-driving cars, it is definitely going to happen; that’s not really up for grabs. The most common retort you get is, “Oh, well, it won’t be as good as me as a driver”. Well, one, you’re probably a bad driver, but, two, it just needs to be better on the average for it to make sense. Yes, your AI is not going to paint like Picasso, but neither do any of us. It just needs to be better than the average.
There’s all these sacred cows people in the AI field used to point to. The most obvious one was this game, Go, which was the last hallmark of humanity’s intellect. This year the human grandmaster at Go was defeated by Google’s AlphaGo, which is basically a lump of code. That’s it, we’re done.
The implication here is if there’s a task that can be expressed to a machine as input and your desired output, then it’ll get near perfect immediately, because computers can learn in seconds what takes us our entire lifetime. They can read libraries while we’re tweeting a single tweet. The best part is they don’t even need to read libraries, they have all their friends who can read libraries, and they can network with them, and they can get all the same information. They have all these cruel advantages that humans will just never have.
The other retort you’ll hear a lot is, “Well, how is it going to learn off me, I’m really good at Photoshop”. Well, did you ever think that when you’re using Photoshop, you might actually be teaching it what it is to do. So what should you do? Nothing, it’s the future, it’s happening, deal with it.
The key idea here is that for many tasks, we’re going to learn that good enough, immediately, for free, beats the old world. Good enough, immediately, for free.
The questions that you should consider are:
- Does AI or machine learning make it cheaper, faster, easier for your customers to make progress in their lives? If so, there’s repercussions for you.
- Does your product have a set of data that a machine could learn from? If so, you need to start learning from it, and if not, you better hope your competitors don’t.
- Can you observe current user interests and behaviors and from that help future users so that they don’t have to jump through the same hoops as the previous people? If you can, it’s going to be a fantastic world for your next users. If you can’t, your next users aren’t showing up for you, they’re showing up somewhere else.
Messaging and bots
A trend that’s really popular at the moment is messaging and bots – they’re connected. The reason there’s a buzz here is because bots are at the intersection of messaging, which is itself a mega trend, and AI, which we just covered. Bots kind of sit in the middle.
What’s going on with messaging is pretty obvious. The internet is being rebuilt around people. We used to always build an internet around pages and destinations and the idea was that me and you go to The New York Times and we both get to see the exact same thing because it’s a page and we’re people. What we’re seeing is a gradual shift. Most products now map themselves to people. Most content sites are learning that when they go there, if Des reads sports, let’s show him sports first. All that stuff is starting to get really obvious.
The other implication here is that people want to be a part of something. It’s a core thing, humans just want to be connected. Every human wants two things, they want to feel unique and they want to feel connected to other people. For as long as there has been computers, we’ve been trying to make people connect through computers. We can see this if we go all the way back. In 1973 Talkomatic was the Slack of its day.
1973 – Talkomatic
They didn’t quite get the five billion dollar valuation, but this was Talkomatic in ’73. Then, in ’81, we had the CompuServe CB simulator.
1981 – CompuServe CB Simulator
This was a closed version of IRC, which is effectively a type of Slack. Because, remember, the jobs people try to do don’t change, the ways you can do them change all the time.
Messaging has finally blown up because of mobile, because of the internet, because of things like push notifications. It’s gotten to a point where you can reach somebody in Australia instantly with a single tap of a button. That happened and we didn’t really notice it, but it’s why messaging is huge.
Above is the top ten messaging apps, and as you can see, even the shit ones you’ve never heard of are growing. That’s how big messaging is. Mark Zuckerberg said recently that messaging is the only thing they’ve found people do more often than social networking.
Which is why we have all these damn messengers, right? Every single product above, with the exception of the bottom ones, is a messenger. Everyone’s chasing it like they want to be the next WhatsApp. They want to get their $19 billion payout.
Messaging has just exploded. Now, what that means is we’re all spending our time on messaging type products. If you’re a product maker, there’s implications there because your users aren’t in your product anymore, they’re probably in Facebook Messenger or in Slack or somewhere else talking to their friends. The product now needs to be a part of the conversation.
If you and me are having a talk about what pizza we should order, or what movie we should go to, that means Domino’s and Fandango needs to be in this conversation, they need to be there with us rather than being a separate destination we go over to and then jump back over to our message and carry on.
That’s what we’re seeing right now and this shift is being sort of described by Chris Messina, who amongst other things, invented the hashtag. He calls it conversational commerce, this idea that commerce can happen as a byproduct of a conversation. If I’m talking about shoes, shoes can appear. Why shouldn’t that be the case? We’re talking about the shoes, we shouldn’t have to go to the shoe place and then come back, that’s not how the real world works.
There are a few interesting shifts here, but one of them is just simple messaging UI. Does it make it cheaper, faster, easier for your customers to make progress in their lives?
We’ll get into that in more depth in part three of this post.
Part Three: Keeping Your Product Alive
We need to talk about simple messaging UI.
Does a messenger make it cheaper, faster and easier for your customers to make progress in their lives? There’s a good chance the answer is yes, and you’ll need to meet them there if your product is to survive and thrive.
This post was delivered as a talk at the San Francisco stop on our 2016 Inside Intercom World Tour.
Finding contextual relevance
Put simply, if use of your product is triggered by a conversation taking place in a different product, maybe you should go to that product and be in that conversation. Let’s say you and I were talking about ordering a pizza, maybe Dominos – or a better pizza company – should be in that conversation too.
One interesting thing that surprises and frustrates me: when I was learning to program we used to build programs like this all the time. You would type in commands and things would spit back and you would say yes or no. Low and behold, you’d be all set. We consider that to be old school today. For some reason when we do something slightly similar through a modern messaging UI, it gets a different reaction. People literally shit their pants when they see this.
I keep asking, “How is that not a command line?” It is just a command line but “it’s the command line for normal humans”, as Sarah Guo calls it. Not for the rest of us – the people who grew up learning the things like zsh. This is for normal people, which is incredibly powerful. But didn’t we move away from the command line for a reason? That’s another question that I consistently ask myself. Shouldn’t we be ashamed? Is this not a failure of UI? Isn’t this just like going through Monkey Island all over again?
Isn’t this just like Monkey Island all over again?
The most common criticism you’ll hear along these lines is, “Well, look at this. Through conversational UI it takes 73 taps to order a pizza.”
Image source: Microsoft Bot Framework
It does. That is correct. And yes, you can order a Domino’s pizza within five taps using their app. I would 100% agree. But that’s not the right way to think about usability.
One button access does not necessarily mean you’ve designed a good UI
It’s not, can everything be done in a button? You should be asking, can everything be learned and discovered quickly by people who want to do it? Not everything is worth learning. Your product can either use a UI, which your users don’t need to learn because they already know it, or you can ask them to learn a new one.
- The question you should ask yourself is, would a familiar UI make it cheaper, faster or easier for your customers to make progress in their lives?
- Other implications include things like, are you asking them to learn a UI for no good reason, or how to navigate your site for no good reason?
- Would customers see any value in knowing how to use your product or knowing how to navigate your site?
- If they won’t, why would they bother? Shouldn’t you just be in that conversation or just communicate with them through chat? That’s what they want.
There’s a difference between using software to sell a product and selling software that is the product.
One area where this may not be true is when you’re actually selling your software, when software is literally what you’re selling. Everyone probably thinks they’re selling software, but there’s a difference between using software to sell a product and selling software that is the product. In the former case, you could be selling shoes and just using some dope React UI. In the latter, you’re actually selling permission to use the product – as in, users want to log into this thing everyday and do their job.
From a customer’s point of view there’s a difference between a customer using software to buy their product, as in using your store to buy their product, and buying software to get their job done. When they want to use your product to get their job done they will want to learn the best UI that makes it most powerful so they can make most progress in their life.
The rest of the time, when they just want the pizza, they don’t want to learn that you’ve got a radio menu to order pizzas or you’ve got some new touchy thing. That’s not what they’re interested in. They would love to say, “Pizza, please. Pepperoni. Done.” Next problem.
From screens to systems
Another interesting trend is this idea that your product is not a set of screens and it’s not a set of user flows, journeys or Jobs-to-be-Done. These are very antiquated ways of building software. It’s best to think of your product as a system. To give you an example of why we talk about systems a lot, I’ve always been really impressed with this diagram that Paul Adams, our VP of Product, drew of Facebook.
He said, “Facebook is a system. There are stories and stories can appear on a News Feed and the Timeline or on a page.” I said, “No, no. But you’ve got Spotify in there. You’ve got all these photos.” He’s like, “Yeah. They’re all just stories. We listen to a song on Spotify, that’s a story, and it appears on a page or it appears in a group.”
I realized he was right. It’s quite stunning how such a complex product is actually such a simple system. I began to think about this more, and one of our researchers drew this diagram to explain Instagram, which I really like. Instagram is a system. You take a picture of a cat, you apply a filter and people can like or comment on that cat.
The real value in thinking in systems is that how, when and where users want to interact with your product will always change. If you’re doing your job right, some of your best customers might not even use your app or your product, which is really weird. This didn’t become clear to me until I met an Intercom customer and we had a conversation that went like this:
Customer: “Hey, love the product.”
Me: “What do you use it for?”
Customer: “Oh, I don’t actually use the product.”
Me: “Oh, right. So is it the blog?”
Customer: “No. I use the product, but I don’t use the product.”
I thought “Oh, no. Not one of these.” But his point did actually made sense. He was using our Facebook integration and replying to customers through Slack, and as a result he never logged into the thing that we have 120 people lovingly working on in Dublin, which will break their heart.
If you think of these things as a system, you can get those use cases that you didn’t really consider. If we thought about Intercom as a screen, plus email notifications and other bits and pieces, then when something like Facebook Messenger for Business launches we’d say, “Uh oh. We need to redesign the entire product again.” When you think of it as a system you’re thinking, “That’s cool. Facebook’s just another way in which people can communicate with customers.” It fits into the system. Now we can receive Tweets, or now we can receive Facebook messages, and we can receive emails, and people can reply through Slack or through email or through push. Whatever. It’s simpler when you have it as a system, because you don’t have to constantly redesign things.
- The last question and implication here is, would making your product usable through all mediums, through all inputs and outputs, make it cheaper, faster or easier for your customers to make progress in their lives?
If it would then you need to think about that because it means that you won’t be surprised by things like when a watch app comes out or Amazon Echo wanting to integrate with you. If you have a system, you don’t have to worry about it because there are lots of different ways things can come into your product.
That does mean we might be looking at a post-app world, and that’s not actually a problem. We’re not app managers. We’re not in this to build apps. We’re in this to build value for customers and help them make progress in their lives.
I think as founders, as support people, as marketers, as product managers, we have an unspoken job. And it’s not a bullet you’ll find on your job description. That job is to be mindful of all the different technological shifts that are happening in the industry.
Are we constantly and perpetually asking ourselves, do these things actually affect us? Can they help our customers make progress in their lives? If we do that then maybe we can stay alive.