Guerilla techniques for growth

Good afternoon, everybody. My name is Nick. Before I start, I have a huge confession to make. I thought I was going to be speaking for an hour so I prepared 160 slides but they only gave me 25 minutes. So, what I'm going to do is I'm going to skip a section at the end, I'm going to talk really, really quickly. And if I can get through each slide in 20 seconds we're back on target again.

So, who am I? Well, my name is Nick and I'm a data scientist which means according to "Harvard Business Review" and "Forbes Magazine" I have the sexiest job of the 21st Century. But I have a confession to make. I have no former training in data science what so ever. I'm actually a rocket scientist and aircraft engineer.

Okay, biography. Why am I here? I started a software company, I sold it to Microsoft. I spent 15 years working at Microsoft, 10 years in the games business. I ran analytics, real networks, did some consulting, and then I reached out to Facebook. I spent six years at Facebook as a data scientist and I'm back working for myself. So, today's topic. What is growth hacking?

Well, the term was first used in 2010 by a chap called Sean Ellis and he said, "A growth hacker is somebody whose true north is growth." I don't know about you, but I don't understand what that means. So, I searched around the internet. There were some other definitions. Some better quotes. Non-traditional rates of growing your traffic, more focus on objective, typically innovative.

These are all, sort of, good stuff but I'm here to tell you there are no silver bullets and, in fact, there are no silver cannonballs. So, anybody who tells you, "We're just gonna build your app and hit the growth hack button," they're talking, sort of, garbage. There's no secret toolbox. It's in a toolset. You need to think about a growth hacker and part of this presentation is how to think like a, sort of, growth hacker.

So, what are the differences between growth hacking and traditional marketing? If you're in an established business you're properly happy with a 10% growth. If you're a startup beginning your app, you're kinda looking for a 10,000% growth and you can't spend 1,000 times as much money. So, you need to come up with creative scalable gorilla-type solutions to, sort of, get marketing for your product.

See if these things resonate with your app. Marketing is more established, longer frequency, don't have an install base, you tend to have a shoestring budget, you've got an untested product, no budgets. Budgets tend to make people lazy and so they think in traditional ways, but the best quote I can think of for a, sort of, growth hacker was something by this, sort of, guy. Winston Churchill he said, you know, "Gentlemen, we have run out of money; now we have to think." So, this is the kind of mindset you want to put yourself into when you're growth hacking.

This is your product. It's the proverbial tree that falls down in the forest. "If you build it, they will come." If you have the greatest product in the world that nobody knows about, it's just the greatest product in the world that nobody knows about. Content is king but distribution is God. You need to get people aware of your product. Visibility is oxygen. We were, kind of, naive. We thought if we built a really good product it would just sell, and that's just not the case. You need to tell people about it.

Here's kinda the problem. This is a live, sort of, query. Best mobile app, what are you gonna get? You get a billion results and the problem is how often do people click here? Being positions 1 through 10 is orders of magnitude better than being 11 through 20 or 21 through 30. Eighty-six percent of all apps downloaded from the app store come from the top 10 charts. Six thousand new app released every day on Android. That's over 2 million a year. So, to put in other words, "You are here." It's pretty depressing.

So, you'll find out that user acquisition costs are probably your single biggest expense. This is cannibal data. Lots of companies, sort of, do it but it's pretty much demoralizing. You're gonna spend an awful lot of money. Are there ways you can, sort of, try and build traffic without having to spend all that money? Well, the difficult part though is that history is written by the victors. You don't hear about mistakes, this is survivorship analysis. Just because something worked for somebody else doesn't mean to think you have to replay those things, but you have to think in their, sort of, kinda mindsets.

I'm not going to tell you to jump up and dance like a fox. Dumb Ways to Die. These things got shared virally. What is it about them that encouraged them to be shared virally? Whether it's Potato Salad, whether it's [inaudible 00:04:00]. All of these, sort of, kinds of memes or themes, you know, how many shares you need to get, a ticket to space, the egg. All these things. These things people shared. Why did they share? Well, it comes down to it's people wanting to share the things, not because they were told to share things.

If you have an app that says, "Hey, share this with all your friends." Why? They don't care about that. It needs to have some reason for why they want to, sort of, share because at the end of the day it's the product. If you have a good restaurant, "We can serve you great food." Not because you tell people you have great food, it's because you actually serve great food. So, growth hacking without a good product has kinda been like a dodgy used car salesman.

If you're not gonna get it right with the first time, you're not gonna get it right with the second time or third time. I spent a lot of, sort of, time in games. Players hate bad games too. If you promote a product that's garbage then you've just damaged the market for everybody. So, the first rule is you've gotta serve great food, you've gotta have great product. So, a dodgy used car salesman, make a great product before you start trying to, sort of, hack it.

The next thing is, sort of, kinda leverage the power of the virality. I'm sure you're all great salespeople and if you can get on the phone and tell someone about your app you could probably sell your app to the, sort of, person but you can't scale. So, you need to get your users to be ambassadors. One person tells two, two people tell four, four people tell eight and so on from there. It scales easily. Leverage somebody else's product because 92% of people will trust recommendations from their friends, which is pretty powerful.

A lot of data science stuff. I'm gonna skip through a lot of this stuff but, you know, you need to be AB testing. If you're not doing AB testing, sort of, kinda go home. It's that important. It's just as important as product design. When I was in the games business, you know, even the difference between play and join. Play kinda sounds fun. Joining, if I wanna join the game it, kind of, applies a bit of a commitment.

Making half a percent in changing something, the denominators are huge. There's hundreds of millions of people, sort of, playing these games so if you make a half percent improvement it makes a huge change. And it's the tastes of the fish, not the tastes of the fisherman. Don't be a Prima donna, you are designing a product for somebody else so data levels all arguments. It's, sort of, as simple as that.

Iteration is the name of the game. It's wash, rinse, repeat. Wash, rinse, repeat. Wash, rinse, repeat. You need to try something. You need to learn from the results and the mistakes and iterate. Wash, rinse, repeat. Wash, rinse, repeat. If you don't measure things you don't know whether they got better or worse. There's no such thing as a failed experiment, only experiments with unexpected outcomes.

A better, sort of, quote. Andrew Lang is a famous statistician. He said, "He uses statistics like a drunken man uses lamp-posts, for support rather than illumination." What does that mean? If you have this, sort of, hypothesis, "Hey, I think if I reduce the price of this I'm gonna sell more," and you reduce the price, you sell more. Well, you've just supported your hypothesis there. Well, the data scientist should be there. [inaudible 00:06:38]. How much should I reduce it by in order to get the increased boost in sales? So, think about how you're gonna leverage these things. "Without data, you're just another person with an opinion." "It's only a mistake if you don't learn from it." Do you know who said that? I think it was somebody's mom or dad. It's not really important, but it's just it's important it's a mistake you don't learn from everything you do.

Data-driven again. Testing beats all guessing. Don't listen to me, listen to this guy, Sherlock Holmes. "I never guess." It's a capital mistake to guess. So, always be testing. And I'll be mentioning this later on because, you know, we measure, we improve. There's a fascination in the industry for measuring, sort of, daily active users, monthly active users, or daily active people as a great measure of success of the app. It goes up and down.

Why does an app, sort of, daily active users go up and down? Well, you know, you lose people. They sort of hemorrhage out of the bottom and then you gain people. If all you're doing is measuring the daily active users of your app, what you're doing is you're hiding the problems behind it. Imagine there's two apps. App number one over here has a million daily active people today and a million daily active people tomorrow. Another app over here. A million daily active people today, a million daily active people tomorrow.

This one over here is garbage, it hemorrhages people. Everybody who tries the app tries it and then, sort of, falls away. But your user acquisition team is so fantastic they're [inaudible 00:08:00] bringing you fresh eyeballs and fresh meat every day. This one is just digital crack, it's cocaine, everybody who tries the app just loves it and, kind of, [inaudible 00:08:07] but the user acquisition team, sort of, sucks and they're not bringing in any fresh eyeballs.

But from just measuring the daily active users you think, "Well, these apps are, kind of, about the same." So, there's different, kind of, problems [inaudible 00:08:16] problems. So, in order to be able to do this, you know, next advice is you need to create, sort of, dashboards for these things. Acquisition, engagement, retention, monitorization. You know? Go out and make dashboards for these today. Stop until you've got this stuff because, you know, the price of light is less than the cost of darkness.

And I'll kinda give you a dating analogy for these things. Anybody can get a first date. First dates are pretty easy. Getting a second date and a third date, that's kinda hard. So, if you're in the first date you're in the acquisition business, you're in the eyeball delivery business, you're bringing in that to the fresh eyeball, that fresh meat into the site. Can you get a second date? This is kinda harder. Only one in six apps get a third date. So, how quickly could you get the user to that user? They've gone through all the effort, you've paid all the money, acquisition costs, and then they'll, sort of, get through there and after, sort of, two or three attempts they're not gonna use the app anymore.

So, your app has to be, sort of, very...get users to have value proposition in the app very, very, sort of, quickly. Getting demoralizing for you now. Twenty percent of, sort of, content is never used. Half of Reddit posts go completely ignored. This is, kind of, I don't have a lot of time to go into this data but what this would tell you is people's phones, they only have, you know, between 10 and 100 apps on the phones. There a few battle [inaudible 00:09:28] apps. If everybody got our phones out, everybody's got Facebook, Instagram, Snapchat, all those, kind of, things. We all have this corset of apps but all the rest of the stuff is spread around lots of different, sort of, thin things.

So, there'are only a very small number of apps and it's a really strict [inaudible 00:09:42]. Again, [inaudible 00:09:43] you can't, sort of, see. But this data represents the top 1% of apps compared to the upper quarter, 25% of apps, and then the top 1% compared to 5%. You'll find that the ones right at the top they're the ones that have orders of magnitude, sort of, better sales, sort of, conversion rates so it's really important to be a top-grossing app.

Relationship. Again, you've got the effort of [inaudible 00:10:03]. Buckets with holes kinda need fixing more than they need more water. And I'll talk about this a little later on. But we're in the second generation of developing apps now. Once upon a time, people say, "Oh, get this [inaudible 00:10:14] prototype, get the product out there. Get users as test cases to try and find out what works, what doesn't work." But that's wrong now because people try your app and if it doesn't work they've just moved onto, sort of, something else so unless you can deliver a perfect experience the first time they try the app you're gonna lose out.

[inaudible 00:10:28]. Monetization, again, gains us a lot of the stuff I've, sort of, gone through. Now, games are free. Once upon a time, people go into shops and buy copies of Flight Simulator and things, they put their $50 down and then they go home and they either enjoy the game or suffer buyer's regret [inaudible 00:10:43] Now, games are free. And in order to encourage people to, sort of, want to pay money you have to deliver a, sort of, great experience. And because they play the game, they play the game, they play the game, and after a week the pay gate comes up, "Well, you know what? I don't mind paying that extra $1 for this particular amount of money."

So, again, your players become, sort of, payers. It's important. So, you need an ongoing relationship. Don't be this guy. If you went on a date with this guy and said, "Well, actually I'd prefer if you shaved next before you come on a date, or then I would have a shower or wear a shirt." If you go on a date he's not gonna get a second date. So, again, if you produce your app saying, "Well, it's kinda this [inaudible 00:11:17] work out what works or not," you've lost the game. "I can change."

And then finally, kinda resurrection. It's cheap to bring someone back from the dead than to find somebody new. Some data. Nordeus, a top football manager, it's a game on Facebook. They spent $3,500 on a re-engagement campaign. "We loved you. We missed you. You used to play a lot. Here, have some free coins. Come back to the game." Overnight, they get a $121,000 return just by spending a small amount of money on a re-engagement campaign. Keep at it. Everything compounds. A little bit of math here. Make a 1% improvement in everything you do every day for 365 days you'll end up with 37 times where you start. Make a 1% decline you'll end up with about 2.5%. You'll just whittle away. It's a glass half full, glass half empty.

Growth hacking isn't new. I remember these things. Why would a company call themselves AAA Plumbing or [inaudible 00:12:08]? Why would there be [inaudible 00:12:09] forms of Heathrow Hilton, the Hilton Heathrow? They were trying to get more [inaudible 00:12:12] on there. When I was at, sort of, Facebook, one of the top games early on was (Lil) Green Patch. Why? Because they had the round bracers and that's sorted ASCII-wise at the top of the list. So, when you searched for list of games it was always at the top of the list.

Listen to Jeff. He called Amazon, Amazon because it starts with A. He wanted a company that appeared first in the listings. You've gotta do everything you can. We are in a, sort of, crazy society, the firehose of, sort of, content. You have Twitter streams, Facebook feeds to, sort of, go through. You have to, sort of, standout some way. You have fractions of a second to get noticed and cause a little bit of friction. So, we need to talk about things, creative headlines, the secret to this.

What is it you'd think of that encourages people to, sort of, click? I'm not telling you to do, sort of, clickbait because that's wrong. I've been going back to the dodgy used car salespeople. If you develop someone to your site and it turns out to be a garbage site then you've ruined the market for everybody else. But if you have got a great product, you come up with aa creative headline, what is it that caused that little bit of friction?

There's a lot of science but there's some luck. But you can bend the luck. When fortune shines on you, you have to be ready to exploit it, again, in days. Marketing campaigns used to be measured in months. Social media growth hacking is measured in hours. And remember this incident, it's the Chilean mining incident. Some guy at Oakley Sunglasses decided, "You know what? When they come out they're gonna want some sunglasses so I'll FedEx them a create of sunglasses." It's estimated that 1 billion people watched this live and so they got $41 million of free air time just by somebody who decided, "Oh, think about that. We should ship them some sunglasses."

But you know, we said milliseconds to go through. This is, kind of, the peak of the traffic for the Chilean miners. If you've got a product related to a Peruvian tree frog and it just happens that NPR is talking about some story that's a Peruvian tree frog, within minutes you have to be, sort of, jumping on the coattails of that story in order to be able to, sort of, get your free visibility riding on the coattails of those things. Even if you stand still.

Paris Hilton got banned from some hotel in Las Vegas. I don't even know what it was. But some creative thinking in another brow of hotels they said, "You know what? Paris Hilton's also banned from our hotel. We don't want her there." And they put their website out there and they put their creative things up there. The Google spiders came along and they picked up this story rather than the original story so when you actually search for the headline, "Which hotel was Paris Hilton banned from?" You actually went to this particular creative one instead.

I'm sure we've seen all these things, you know, what particular viral moment can you, sort of, get on top with straight away? IKEA. They improved the response rates of their marketing campaign by 34%. Do you know what they did? They emailed people when it was sunny outside to buy patio furniture. That's, kind of, the way that growth hacking needs to think. Google, they switched the shade of blue on their campaign and they made an extra $200 million a year. If I wanted I could drop the mic over there. $200 million just by changing the shades of blue.

American Express, you know, they had to spend a lot of money. They estimated $100 million of marketing just by putting things like [inaudible 00:15:14] on their cards. There's a lot of ingredients into, sort of, growth hacking and I don't have, sort of, time to go into all those things but lots and lots and lots of different, sort of, ways to get promoting a product. Facebook, my previous employer. It wouldn't be without presentation without doing a quick, sort of, pitch. There are 2.25 billion people now in the world that use, sort of, Facebook so it's a big site.

What logo stands out there? You kinda have to, sort of, get noticed any way you can. Anybody play Angry Birds, the logo of the red bird on there? Why do they use a red bird? Well, at the time they launched the app most of the games were blue and people are sliding through. You have fractions of a millisecond to, sort of, go through and create a little bit of friction. If you're launching a game, which icon of these would you use to launch a kid's game?

No. Test, don't guess. You're probably right, you know, maybe people like three-legged dogs but at the end of the day, data levels all arguments so you have to, sort of, test these things. Although you're probably right. Eyes seem to, sort of, sell quite a lot so if you're producing a game, the bigger the eyes, the bigger the puppy dog eyes the more likely you are to, sort of, get sales to go through. But look at all of these slides. Instead of looking at the slide, you can tell the play mechanics straight away. That's a match-three. This is a farming game that's, sort of, gone through.

User's sliding through. Slide, slide, slide, slide, slide, slide, slide. And again, a couple of hundred million people very large denominator. If you make half a percent change in conversion of people clicking through to your game by testing these out. So, you should be AB testing every single slideshow, every single screen, every single icon that, sort of, goes through.

Even choosing, sort of, colors. If you're into games, this is the color palette and a business palette. Which one would you, sort of, come through? In people's minds. Fuzzy trace theorem. Peoples minds remember things in, sort of, different ways. We have two ways. We have something called a short-term recall and a long-term recall. If you go home and I ask you, "What do you remember about this presentation?" You might remember specific details if you walk out straight away but when you go home you think, "Well, the guy talked really fast. He was bloody brilliant but I can't remember anything he, sort of, said other than the gist of what's going on."

Let's say you go to a comedy club and see a comedian you go home and talk to your spouse. "Wow, he was a really funny comedian. He was hilarious." "Well, tell me one of the jokes." "Well, I can't remember any of the punchlines, I can't remember any of the jokes, but I remember he was really funny." So, fuzzy traces how important it is to, sort of, get people to share your apps and things. Going back to the app icon. "I doubled my app installs by changing my icon." Boom, and he doubled his revenue. What he did? Well, he changed one icon to the other. It's kind of obvious, you know, it's a jigsaw puzzle game, the icon on the right tells you it's a jigsaw puzzle game.

But again, fractions of seconds going through I'm looking for a jigsaw puzzle game. Why would you have clicked on the one on the left beforehand? The world is moving mobile. We check out our phones lots of times a day. We, kind of, already know that. U.S. adults will spend over 3.5 hours a day on their mobile devices. When did this revolution start? It's, kind of, when the iPhone, sort of, came out. We all know those things. But [inaudible 00:18:10] data. It is not from cannibalizing desktop usage. The fact that we're using more of our phones. I still have my laptop, I still work at my desk and I use my mobile phone for other things.

So, those are additional hours that are coming through they're using it at the detriment of other things. People aren't watching as much live TV anymore, they're not going out to the movies as much more, they're not playing with their kids as much anymore. That's, sort of, sad all those things. But it's not cannibalizing, sort of, desktops. So, the day of the desktop PC is still around. Why? If you look at mobile phones you're all carrying in our pockets more processing power than what sent men to the moon. It's pretty staggering.

There are five billion mobile phones in the world now, about 60% of them are smartphones. So, another jaw-drop moment. Over half of humanity now has a smartphone. They're used for lots of things. You know, my smartphone never leaves my side, the first thing I do. Forty-five percent of people take photos at least once a day. Ninety-percent of them share once a week. Seventy-percent of them online. Snap, snap, snap. People use their phones for navigation. There's no surprise. People use them for following news. Ninety-percent of time spent on mobile device is spent on apps and 58% of mobile purchases.

And that sounds like, you know, you should go make apps. But I have to put a caveat on there that over 50% of that is spent in Facebook alone and 90% is spent in the top five apps. So, if you're just starting out there, again, it's a pretty brutal marketplace. If you're not really sure, don't go to the app first, go to the website version or the mobile version of it to, sort of, go through. Okay, I promised to let you know why people share. Kind of sharing is new. It's just digital gossip. For years, people have always wanted to share things and talk about what, sort of, goes on.

But sharing comes from the top of Maslow's Pyramid. It, sort of, goes through self-actualization, realization. These are all [inaudible 00:20:06] what, sort of, people share. We don't have a lot of time but there's various different, sort of, content of things. There's content creators, content aggregators, content consumers. I'm sure we've all watched a YouTube video. How many have actually created a YouTube video? Not as many and it's a, kind of, pyramid but there's aggregators. There's plenty of people who said, "Well, you should check this out. Here are some great videos I found that may still find..." So, there's a pyramid it, sort of, goes through. So, depending on whereabouts you are on the pyramid there's different techniques.

It kinda sounds obvious but people share things when it's really easy. Don't make it hard. If you don't have a Share button or a Like button on your app, you've just kinda missed out and people make it easy to go ahead and, sort of, share things. Sharing your social currency. We share things that makes us look cool. "Look what I found. Check out this funny video. You're not gonna believe this. Look at my house. I'm gonna be the first to post this. It'll make me look cool." It's, kind of, that sort of feeling. It feeds that, sort of, dopamine, sort of, cycle that makes us, sort of, feel good.

Triggers. People share things when they're reminded about stuff. "Oh, that reminds me about something I saw." So, if you're picking a game, base it on a current theme. Anybody know who this is? Friday. When do you think people search for the word Friday or Rebecca on Google? When do you think people search for a full moon? If you look at the data, people search for full moon when there's a full moon outside. So, if you've got a game or an application that has triggered on something it's more likely for people to share.

People love sharing things related to emotion or excitement, laughter, and anger. But surprisingly, not sadness. Check out this movie. Here is some, sort of, data about more likely things being shared for "New York Times." So, people don't share sad things but they'll share other things. And also related to emotion if you look at how well campaigns are gonna do well. Emotional campaigns are much more likely to be better than, sort of, data-driven ones.

Practical value. People share things when they think they're gonna get value from. "I think you'll find this useful. It'll be perfect to help you with things." Coupon sharing. People share stories. "I called their support line." Nobody says, you know, just, "Verizon sucks." They say, you know, "Verizon sucks because let me tell you a story that, sort of, goes through." I hope there's nobody from Verizon in the audience. People share because others do. I know, sad bear here but if you don't share things, maybe nobody knows you're there.

Some people share because it's a tribal thing. We share things with people we're affiliated with. The soapbox. People share to broadcast what they believe and they try and influence other people. But everyone talks about the, sort of, content but that's, kind of, all the data shows that that's actually not true. Surprisingly, many people attempt to understand things before they share. "New York Times" did a, sort of, survey. Seventy-three percent of people said they process information more deeply when they are about to share it. "Well, I think people are gonna judge me based on what's, sort of, going on."

And again, 85% of people said that reading other people's responses and then post this to them helped them understand and process the event. So, it's a good way of actually, sort of, validate it. We think, "Hey, I've just read this article that says this. What do you think?" And you post it so people [inaudible 00:23:11]. People share to reinforce the image about they want to present about themselves. We kind of know about those things. Sometimes people share for help. Some people share just to remain connected. It's just an excuse to communicate. Back to the gossip.

And others have an obscure hobby and it's just an excuse to share things with people that have an interesting, sort of, hobby with them. But, kind of, the most important message about all this, sort of, stuff is sharing's how customers connect with each other, not with you. So, if you have a brand that said, "Hey, share my brand just because I have a brand." People don't care about that. They share because of how it makes them feel and they wanna communicate with other people they like, not with themselves. This is the section I'm gonna have to skip so I'll, sort of, go through all of those things because I've got one minute left and I wanted to get to the conclusion slide. This is just data showing how we can bend people's minds with how you share data and stuff.

Click, click, click, click, click, click, click. Click, click, click, click, click, click, click. Click, click, click, click, click, click, click, click, click, click, click, click, click. So, you need to do some growth hacking. How do I find someone on my team to be a really good growth hacker? And this is a slide I'm, sort of, gonna end in. You need to search for somebody who loves data and understands it, somebody who's a little bit curious, a little naughty, "What can I get away with? What would happen if I tried this?" Somebody's who's gonna experiment McGeyver side of things, leverages opportunities, moves fast. This is really, really important. You've got minutes to, sort of, get things going in the social world, create scalable solutions and seize all chances of visibility.

You need to grow a culture [inaudible 00:24:37] and as we know, data scientists are the sexiest job of the 21st Century. The best quote I've got about growth hacking is, you know, you're looking for people with Swiss Army penknives, not unicorns. So, your job title says you're a growth hacker. If you're not breaking down walls, you're kinda doing it wrong. So, it's not really about unicorns, it's about being Swiss Army penknives. And with that, I have 30 seconds left and I've got to the end. Thank you very much.