October 16, 2018
Ep. #6, Customer Reliability Engineering with Google’s Liz Fong-Jones
In episode 6 of O11ycast, Charity and Rachel talk to Google developer advocate Liz Fong-Jones about ways to build systems to be more transpa...
Hi, my name is Cameron. Today we're going to be speaking about how to create killer trend stories using data that's inherently part of your business.
A few things about myself, I help young companies that are facing developers grow very quickly in a 24-month period. I've worked with two companies in the past four years. The first one is a company called Redis Labs. We provide a fully-managed Redis as a service. We've raised a total of $20 million in series B funding.
I've led our marketing operations as VP Marketing for the past two years. Under that, I've helped us to grow to more than 5,000 paying customers over an 18-month period. And it's completely SaaS based.
And before that, I worked with a company called Newvem. I was also an investor in this company. What Newvem does, or what it was doing before it was acquired, was analyzing Amazon cost and usage, much as Trusted Advisor does within the Amazon Cloud. We raised a total of $4 million in series A funding, and we were acquired by a large MSP called Datapipe.
So what you see here is the result of a trend story at Newvem after being in the market for six months. The challenge that we had was we made a funding announcement, we made a G announcement and we're left with very little to tell the public media.
After this story, and I'll give the example in a few minutes, we had more than 1,000 unique visitors within 72 hours. It was featured on VentureBeat, and even Guy Kawasaki mentioned it and also retweeted to social media feed. And once we pushed this to our blog post, as a blog post it achieved more than 10% of users over 30 days and was one of the top 10 pieces of content in our inbound marketing until we were acquired.
Why trend stories? Why are these important? These days, the objective, if you're trying to market yourself to developers, is to earn their trust.
The best way to do this is to literally surround the IT organization with content about the problems that you're trying to solve. It doesn't matter if the organization is one person, if it's 10 people or 100 people. At this stage in the game, developers are so suspicious.
I put this image of this cat for a very specific reason, because a developer is suspicious, isn't very trusting of brands, and is literally, as I've always kind of fashioned it, like a little cat in your house that is a bit timid and comes out and will scratch you if you get too close to him or her.
There are great facts to substantiate this. Forbes says that 67% of the IT-buying journey is made with research. So this means that when you're at a stage where you're selling your offering, if you get an inbound lead, by the time a salesperson speaks to that lead, they already know everything about your offering.
They're already speaking to your competitors, and they're already armed with enough research that they can have a real discussion. There isn't much education left when you speak with a hot lead.
Secondly, if you are an early-stage company, just raised series A or series B funds, there's no way you're going to outbuy your way through. You're not Oracle. It's not going to happen.
Number three is if you did go so far as to make a funding announcement, or you made a G announcement, or made product announcements, these are very great. But it's very hard to get that consistent coverage from influencers and from bloggers that are interfacing with the people you're trying to engage.
The fourth one is there's a lot more to your business, if you think about it, than just offering a service. Because you're a specialist in the problems that you're solving. You can literally transform a lot of the data that you have, you probably don't even know about it, and become an expert and literally a news channel that will feed this news to people who want to see it.
The fifth thing, and probably the most important thing is that when you make the right kind of story and you promote it correctly, it delivers itself time over time. We still have articles, actually at Redis Labs, you may not even believe it, but we still have content that was published months back, and bloggers are still referencing this content today.
So when it comes to creating trend stories, there's really two different approaches. The first one is hijacking stories. And this literally means, let's say for example, that you're a monitoring company. If you see a lot of coverage from larger players, the idea is to be a bit reactive, right?
It's to reach out to those blogs and say, "Hey you know, this is what I'm doing. We're relevant as well." This is a very difficult game to play. It requires a lot of resources and quite a bit of dedication.
The second one, and what I'm going to focus on today, is curating content. That's about discovering content that's behind the scenes, creating it in a way that it's very easy for a blogger influencer to understand, and to promote it.
There's three forms that we'll be discussing, first one is customer data. This is literally the data that you see within your system that can substantiate a very interesting claim. The second one is using lighthouse stories from some of your best customers to tell your story better than you can. And the third one is using user behavior of the community that you're engaged with.
Let's speak about getting customer data. So imagine you have a few thousand customers, it could be free users. Why is the information that they have inside of your system important? The first one is, you may not know it, but you actually are already an expert in the domain that you're facing.
A good example of that is, at Newvem, we were experts in how users were using the Amazon Cloud. They used our service because they wanted to reduce costs and improve their usage of Amazon, but we were able to see what they were doing behind the scenes: utilization, availability, cost and security. We literally had this awesome treasure trove of data that an analyst or blogger would die to see.
The second one is analysts and writers. They love this stuff. Because, as an expert in this space, they don't want to hear an opinion. They don't want to hear someone's perspective. They want to see real numbers and real facts to substantiate why that community actually is relevant.
The third thing is if you create content and you produce it right, and you do it in a very concise way, it's very easy for a blogger to digest and, ideally, to promote. So the main message here is that there's much more to customers than just looking at them as dollar bills.
Let's look at an example of that. This actually goes back to the example that I started off with. So six months, you know, great, we've got this great visibility. We're growing, but we don't see a consistent amount of visibility in the bloggers that we're trying to engage. It's gone.
We're just a startup, at the time we were based in Tel Aviv. We're off the map. So we had been thinking, "How can we show some kind of expertise in something that's important in the Amazon Cloud today?" We saw, at the time, there was a lot of buzz about reserved instances, which I'm sure you guys know about, so I won't go into that.
But no one really knew what it is, so there was lots of talking heads. Like, "Reserved instances are good." "They're bad." "The reserved instance is like the death of the Amazon Cloud." "It's the birth of the Amazon cloud." It's just crazy!
There's too much talk about it. So we decided, "Let's take a look at the data of our users, that are using reserved instances, and see what they're using it for." We created a cool report which I'll show in just a second. We made it into an infographic, which an infographic could be debated today, but at the time it was an easily-digestible piece of content for developers in the space.
We fed this to a writer at VentureBeat, and he took it like crazy. It was amazing. It was retweeted across the social space, including Guy Kawasaki. And the end result of this, and the end result of any kind of activity like this, is going back to surrounding your end user with content.
We established ourselves as an expert in reserved instances. And as a result of this, for the next 10-12 months before we were acquired, I would still get phone calls and emails from analysts and writers that were saying, "Hey Cameron. I want to know more about reserved instances. What do you guys see?"
Let's go into some steps that you can easily execute. Now this is not a prescription, all right? It depends on how your business works. But here's a few guidelines on how to come up with a great story based on customer data.
First one is take a very deep dive into what your customers are doing, and take a look at the full funnel from the time that you see them coming to your site, where they're coming from, when they come, all the way to when they try, when they buy and when they leave.
Go very deep into this and see if there's any data that you can match and latch onto for a very popular trend.
Second thing is when you take a lot of this data, and your sample is statistically referenceable, you can essentially anonymize the data. That means you're not speaking about what a customer did with this data. The third thing is create a very simple report. This is an example of an infographic. I mean, it looks kind of crazy, but what you can't see behind this infographic was the pitch that we gave to the bloggers, and it was very simple.
We took three main takeaways. Right? Takeaway one, two, three, do you want to see more? See below, and we would list out all the data points that we had. This created a response, not necessarily a yes, but at least a means for us to make a discussion with that blogger.
Once you identify the top-tier bloggers you want to engage, find a lighthouse blogger or influencer that will lead in a break-in, stratifying the one with the largest microphone. At the time it was a cloud writer at VentureBeat. Then take it to your social channels, take it to secondary influencers. Once you get this initial burst, you can then take it to your own blog, post it as a blog post, and push it out to your community and your leads.
The second one is a customer story. Now, it's not a case study. Case studies are great, but what a customer story really is about is figuring out how you can ride, how you can make their success your success. Why is this important? Bloggers and analysts don't really want to hear about you. They don't want to hear from you. It's annoying. It's like vendor blah-blah.
Your customers will tell you the best stories, especially when you find high-visibility ones.
Second one is no one cares about you after your funding announcement. No one's going to listen to you. No one's going to open up the emails. Nada. The third one is when you can solve the customer's end challenge, and not like an infrastructure challenge.
But think about how you could help a large consumer-facing application, to solve a challenge of how they were able to keep their shopping cart up-time by 99%, or how they're able to win Black Friday, or Cyber Monday, and I'll talk about an example about how we did that exactly in a moment.
The fourth thing, and it's probably one of the biggest secrets that I think is in this space, is all of the marketing PR efforts, in any of these companies, is all pushed to their main offering. It's not pushed to their R&D team. So the R&D team is actually, in PR currency, very poor. They don't have a spotlight onto the technology, and as a result, they can't create a nice pipeline of talented developers to join the company.
Let's see an example of that. Redis Labs last year, it was November. Just came back from Amazon re:Invent, boring time, right? Can't do anything with these leads. Bloggers are going home. The whole business is, not our business, but everything just seems to be coming down to a slump. So this is a great opportunity for us to look at what's happening right now and ride on that with a really easily understood story.
We thought a lot of people who use Redis are ecommerce sites and travel sites, and they're facing huge challenges with Black Friday and Cyber Monday. So I thought, "Let's create a story on how a great, well-known ecommerce or travel company really won Black Friday and Cyber Monday, and Redis was a key part of that win."
I shopped around. One of our lighthouse customers is Hotel Tonight. CTO was like, "Yeah, this is kind of interesting to me." I said, "This is a great opportunity for you, not just to, you know, give some love back to Hotel Tonight, but it's also a great way to give love back to how you actually built the technology in the stack out. Not to mention we're going to put this in front of some awesome developers that are going to see that, and that could turn into some nice developer pipeline for you."
Created a nice story, said, "Yeah, as a matter of fact, we had a pretty big load for a promotion that we ran, and thank God for Redis." "This is great, let's run with that." So the story was very simple, you know, how Hotel Tonight, if you can't read that, it's how Hotel Tonight handles the demand for instant booking and check-ins.
He featured a challenge that they had which was we had this seven-dollar reservation deal, and traffic increased eightfold, and a key component of how we were able to handle that was Redis. Writer took that, made a great story and told the story about Redis better than we could as a vendor.
In this article we had eight mentions of Redis, and he literally said, "It's just a must-have part of your stack." It was great.
We were able to get in that discussion about how a very specific customer type is using our solution to win.
One important note that I want to add here is many cases, if you're fortunate enough on TechCrunch or VentureBeat, or some of these high-visibility blogs, you may not get linked anymore to your site. The days of expecting huge waterfalls because they link you are officially, effectively over with.
You might get lucky, but it doesn't always happen. So many cases they'll link back to the CrunchBase page. They might link back to a blog post. They might link back, in this case they even put some links to open-source Redis, which is still very good. The goal here is you want to be part of that discussion, and again, to surround your user base with content.
How can you do this? One is find customers that you have that are celebrity customers. It doesn't matter if it's consumer-facing companies or B2B-facing companies, find them in your own customer base. If you don't have them as customers, find them as leads. If you can't find them as leads, find them in your respective communities.
Go on Twitter, look for them. And even if you have to onboard them to bring them as part of the story, probably will be worth it for you. The second thing in what really separates this, from being a practical case study into a real customer story, is you've got to find timeliness in the industry.
Forget about going to the typical developer-facing bloggers. You need to go to consumer-facing bloggers in the industry that you're trying to engage.
In this case we were not really touching so much developer-facing bloggers. We're touching guys that are covering B-to-C startups, travel. Number three is create a very simply, easily-digestible story that a blogger can easily look at, look at like three main takeaways, and tell you yes or no.
Four is pitching to the number one R&D challenges. Which is they can't find developers. I need them in my company. Or I found the developers but I cannot keep them focused on what I really want them to do. And number five is promote it very aggressively with your tier-one bloggers you want to engage. And then circulate it in the same way as we discussed.
The last one is about user behavior. User behavior is going outside of your customer box, and taking a look at the community you're trying to engage, and making sense about what's important to them outside of how your offering impacts them.
Why is this important? First, it's actually fun and it's fairly easy to collect because you're suddenly having a discussion with a lead that isn't really about your offering. It's actually about the general problems that they're suffering from. The second one is it surprisingly builds great rapport with prospective customers and also your current customers.
Actually, surprisingly, people like to discuss their general challenges. And when you come in as a consultant, it's very easy to establish a meaningful discussion.
Number three, it creates a sense of priority to get a little bit deeper with your current customers and understand how you can take some of that feedback and incorporate it into your product roadmap. I put this image here for a specific reason.
This is a screenshot of people who are just sitting at Amazon re:Invent. Most cases, when you're a vendor and you're sponsoring the event, I've sponsored at every one of the events, it's very easy to look at all of these guys as leads that will eventually buy my offer.
But in reality, what you have here are really impressive people that know how to use the Amazon Cloud and either paid or had someone pay for them to go to this event, and therefore they have a tremendous amount of knowledge you can actually benefit from.
Let's look at a real example of this in action. It was November last year. The Hotel Tonight story came out, nice. Again, dead. It's November, right? We came back from the Amazon event, and you know, practically can't do anything until the beginning of the year because it's November/December.
So I looked at the leads that we had generated, and I thought, "You know, one of the challenges that we have is Redis is an In-Memory NoSQL database, and it's a fairly new one, so the discussions tend to be around that, In-Memory databases, NoSQL databases."
The challenge was how can I get into a discussion that is much bigger, that's out of that? So I took a look at these leads that we've collected. I'm like, "These are some of the most pristine Amazon users. Let's run a survey and see how they run databases on Amazon Cloud, and therefore it will help us to establish some type of expertise in how to deploy a managed database on the Amazon Cloud."
So we created 10 questions, 10 very simple questions, all mandatory, just clicking buttons, drop-down boxes and I'll show an example in just a moment. We took a very aggressive approach in that I wanted to see results, and I wanted them to be specifically referenceable.
You get a survey result and you have 10 people respond to it. Who cares? The end result here is we emailed this to roughly 1,000 leads, and we had more than a hundred actually respond to the survey, which is a very high conversion rate.
The reasoning for that is we kept a very focused effort. No one wants to complete a survey. So what we do is we gave an offer that developers love, which are drones, a chance to win an awesome DJI Phantom II drone. Someone is going to win in 72 hours. Complete this in 72 hours.
Email one goes out on hour zero, and email two goes out after 48 hours, and the winner was drawn. 72 hours, the story was finished, and the data was collected. Suddenly we could actually join discussions that a lot of influencers were making out of, I don't know, in spite of their own knowledge.
The one thing that personally irritated me is I would see some of these stories that are like, "NoSQL is the death of SQL" and, "Relational databases are here to stay," and "NoSQL is only used for some stupid use case." What the data that we collected proved is that with hard data we showed that developers have a very wide variety of databases. And, actually, the most popular database is MySQL, and they use Redis, Cassandra, Mongo, all in conjunction.
Surprisingly enough, the databases are usually used to support a very specific use case. So we mapped it out, we had numbers to prove it, and we took this PR firm that we're working with, we took it over to VentureBeat. Most of these examples use VentureBeat but it's not inclusive to VentureBeat. He took it and within, I think it was around 48 hours, it was up. And it was awesome.
You wouldn't believe it, but even today, I still see that this data is actually referenced in a lot of talk about what database is right for specific use cases. So how can you do that? This is probably the most interesting one, probably the one that, at your fingertips, you can do today.
First of all, challenge yourself by trying to find a wider topic. I'll go back to the example of a monitoring application. You can, if you stay in the bucket of monitoring-specific, I don't know, topic, interesting. But if you want to be part of a wider discussion, try to find something that's pertinent today, and zero in on that.
Number two is create a very simple survey. I've always stuck with 10 questions tops. Each of them are mandatory. And each of them is multiple choice drop-down box.
There's no probes here, right? I would like the end user to complete this survey, in the same way that I would, which is like, one minute maximum. I don't have time for this stuff.
Test the questions with your developer consiglieri. The developer consiglieri are your top-10, closest developers. Maybe they participated in your beta program. They're most responsive to you. Test them out, and make sure that all the questions that you're asking are relevant.
The design. The design, after testing this, and my assumption always was true, is that the design should look like it was totally effortless. In this case we used a template from SurveyMonkey. We didn't use any investment into that, something very very simple. And definitely just on one page. Nothing that you have to click through to complete the survey, something super super easy.
Irresistible incentive, find what that is. Maybe drones are too old now. I have no idea. But they always worked. Make it worth it. Who cares about winning $50? I want to win an awesome drone.
Something that's really meaningful. As a result of this, you can literally become a news source. One thing that's really important to mention about this is, if it works, if it gets coverage and it gets interest, you can take the same kind of survey and you can replicate it on a quarterly, biannual and, even, annual basis.
You're literally giving reports every quarter about the stage of X topic. Now you've made these great efforts to get coverage, what else can you do with the efforts that you've already invested your time into?The end result is just you have really good collateral that can be applied in different parts of your business, depending upon the stage.
Probably the most explicit one is you can use this and you can implement this as part of your sales channel. Think about, "How I can put this into a nurturing campaign?" Sales guys are notoriously coin-operated. And this is actually a really great icebreaker. If you have a rapport, like some of our sales guys will call up and say, "Hey, you know, we just did a report on how databases are proficient over the Amazon Cloud. Do you want it?"
It's a great way to start a discussion. The definitely-great pieces of content that you can push to your community in your newsletter. If you're working with an open-source community, it's a great way to help grow that community as well. Adding it within social profiles and signatures is also a great thing.
Another thing I wanted to add here is as a recruiting tool it's also very powerful because it helps you to establish some type of precedence, credibility within the domain that you're engaging.
That's it. My name is Cameron. You can reach me on Twitter at @cameronperon. Feel free to open up a discussion on that. I love to chat on that. If you have any questions, please email me at this email address. I'm currently phasing out of my role as VP Marketing at Redis Labs. And I'm looking into new opportunities to partner with talented guys and do something pretty amazing.
Yeah, that's a really good question. The short answer is just find the ones that get the most buzz. In our case, Redis is part of a stack for most B2C-facing companies. It's almost impossible not to use Redis.
Naturally, it's very easy for us to look for unicorn, or unicorn-close companies that were there. So a good way to quantify that is cross-reference them like on CrunchBase. See how much they raised. Look at Alexa rankings.
If you have a huge list, and you don't have time for that, just outsource it on Odesk.
It gives you a guess on how much Web traffic they're engaged in. If that's important to you, it can kind of give you a sense of how well your application or your solution is supporting that, right?
For example, Hotel Tonight was a good one for us at the time because I had identified them as something I use. For me that was a pretty easy one. And they're really hot right now. It was pretty natural for me to go over there and start this discussion.
Depending upon what you're doing, you might be solving a problem that they don't want to have any exposure to, right? So it's always important to find out how their success was your success. The questions that I always ask customers like this is how they solved a problem related to a trend that I saw outside. So for ecommerce, a pretty simple one was, you know, Black Friday, Cyber Monday. That was a pretty easy one.
If you're working with something that's really tricky like, let's say something that's related to security, that's a very tough one because in many cases, people don't want to have visibility to problems that shouldn't have happened in the first place.
I would always try to frame the question is such a way where you help them to solve a problem, like on a new project, as opposed to solving something that was a failure. Like how you helped them to improve a specific use case. This is how I would try to do it.
Many of the customers that we worked with, even for case studies, they simply don't want the exposure, right? It's like a cultural thing for them. But I always go back to like, "Alright, let's talk about how maybe we helped you with something very simple or something that wasn't as business-pertinent, or maybe it was for a new project.
And yeah, if a blogger doesn't want it, I'm a firm believer in you give people what they want, not what they need. So if they don't want it, then you've got to go back and find someone else or frame the story better.
Oh, I don't like that. Yeah, yeah, this could definitely be debated. I hear people doing this. I think this is a huge mistake. I've never done that. Fortunately I've never done that, and I've always found this is really ugly.
It depends what size the company is. Let's say that you find a really awesome Fortune-500 company, unicorn-like company. The thing is if you close a deal like that, it usually has to go through several different layers within this company you don't really have any touch, with like the PR or legal, and this can get really ugly.
And then what? I mean, then if he doesn't agree, you have to say, "Oh well, you signed on the contract?" I mean, this is a stupid discussion. It can work. But I really want to try to find a way that I helped the customer to win. I helped not the business, but the specific person that's using the solution, how I helped them to win. And I really want them to reflect on the story.
Literally I want to make him famous. This is the business proposition. I'm going to help you to make your stack and your work famous because your PR firm is never going to do that. And, as a result of this, I'm going to help you find probably some great developers that will be inspired by what you've done.
Yeah, yeah. That's a good question. I think it really depends how you phrase it, right? At Newvem we analyzed four main categories at Amazon. One was cost, availability, usage and security. Security was really interesting because we had crazy insights. We showed that something like more than half of the Amazon Clouds we were analyzing had a security vulnerability of some kind, right? It was crazy.
We didn't when we took this to the press. But the reason why we did so is we want to expose some of the problems that people are facing and show you how you can remedy these problems by taking these kinds of steps, or you can use our offering to help you resolve that. That's how we went about doing that.
I never encountered any kind of negative feedback about that. On the contrary, I had people that were like, "Thank you for letting me know about that." Or suddenly we would see an increase in the amount of secure measures that many of our customers took.
I think it really is how you phrase that up. You definitely can not share how, you know, a specific customer did it, but I think it's actually very informative. Especially if you're in a specific niche.
I'll use the example of an ecommerce. But if you see a category as part of your business that's very popular, and you share best practices, you share the problem and best practice in how to resolve that, and by the way, I can do it for you, too.
I think it's a great win. The real problem, though, actually, on that is how the blogger, or the influencer, or whoever's going to publish that information externally, how they decide to message it out. And you don't really have that much control over that.
I don't know, some cases this could be manipulated, but I haven't really encountered a problem with that in the past where it came out as being creepy or there's some kind of negative repercussions with that.
Yeah. Oh, that's a good question. Yeah, right? What's the point of all this stuff, right? So the answer is that if you're a good marketing person, you answer this by looking at the influence that that has on your sales flow. I think that the days of getting one-offs, I've seen it, we've done it in the past, where we get great funding announcement or great trend story, and we get some really awesome early adapters from great companies that will try and buy, right?
But when you look at any kind of conversion funnel, I think it's really critical that you map out the metrics that are important to you. And this is really far down the pipe than trying to close a deal from that. So I always go back to showing how we were able to surround this IT organization with content.
You can even apply a share voice metric as the good PR firm will try to do. But it's really with this education that I've seen over time an increase, when we see the increase of visibility that we get from external media like blog posts, on blogs outside, and the amount of organic traffic that comes to the homepage, and the amount of free users and paid users.
I see a very positive correlation, and usually an explosive growth in the last two. So the short answer is these all work together. It is not like I will invest my time into a blog post that will be published on TechCrunch, or VentureBeat or some other site, therefore, I will get paying customers. I don't think it's realistic.
All right, cool. Thanks guys.