Some of the top lessons you’ll learn from the interview are:
- How Digital Telepathy decides which projects to work on.
- What they do to figure out what signal’s are worth following up with.
- Why they involve customers in every step of the process.
- The difference between self-reported and observed data.
- Digital Telepathy’s belief in constant “betterment” and how they achieve it.
Joe: Hi, Jason, I’m really excited to have this opportunity to talk to you about product development and all of the exciting things you’re working on for Digital Telepathy.
Jason: Thanks for having me.
Joe: No problem. In order to get us started, could you give some background about your experience and what you’re working on right now?
Jason: Sure, yeah. I work on the products division at Digital Telepathy. Essentially we make products that empower website owners to integrate good design into their sites and enhance the user experience of their site. My background is that I spent about 10 years in advertising on the interactive side, and I came over to DT because I saw what they were trying to do, which is to really apply good design for the whole web, not just for our clients, but to try and really improve the design of the web for everybody.
Joe: Along those lines for the product you’re currently working on, what’s the general approach that you guys take?
Jason: The biggest effort that we’re focusing on right now is called Filament. We’re still working on the exact positioning of it, but essentially what we’re trying to do is to empower website owners with insights as to how to improve the experience at their sites. We basically look at the analytics of your site and try to automatically surface insights that will help you. So, “Hey, you’ve got a traffic spike on this page. You might want to install this subscription form on that page to try to get and more email subscribers.” Or if it says a lot of sharing is happening on a particular post, we shoot you an alert to say, “Hey, there’s a lot of activity happening here.” If it’s centered around, say, Twitter, then: “You may want to reach out to these top followers and engage with them to keep the conversation going.”
That’s all through a platform that we built called Filament. We’re building the analytic piece of it, but right now we have apps that can help you get more shares or get more followers, that kind of thing. All of this really grew out of an organic need. We have a lot of our own websites. We’re also working with our own clients as well. We just found that we had this need to be able to easily install apps onto sites without bugging our developers or interrupting their workflow. We’re very performance oriented, so we’re constantly wondering, “Hey how’s this website doing? How is sharing doing on Twitter on our own blog? How are conversion rates from one-step to another on one of our own sites doing?”
Answering those questions takes a long time. It’s tough to do. You need to have expertise in web analytics to definitively answer those questions. What we are trying to do is build a system that automatically does that for us so that we can spend more time figuring out solutions as opposed to figuring out what’s going on. It’s almost like keeping a regular pulse on what’s happening on all of our sites.
Joe: That makes a lot of sense. So Filament is a way to automate the analytics process and surface some of the most important metrics for different businesses? Is that correct?
Jason: Right. We’re geeks over here so we’re thinking about it like artificial intelligence for your website. I don’t know if your readers have seen the Iron Man movie where his suit is talking to him and it’s got Jarvis’ voice saying, “Your power level is at 67%, sir.” We want that kind of automation level happening so we don’t have to go figure that stuff out for ourselves.
Joe: That’s really cool. I’ve paid attention to, not just Filament, but a few other products that Digital Telepathy has developed. Is there a common theme as far as the initial development process and how you decide to work on one project or another? I’m sure you guys have plenty of ideas. What’s that genesis period like, and how do you guys come up with an idea and decide what to work on?
Jason: That’s a great question. Yes, we have tons of ideas. Our culture is very much built around the idea of betterment. So making small, tangible, and long lasting improvements every single day, over a period of time, when you look back you can see a lot of the betterment kind of stacking upon itself. We’re constantly looking for that in terms of the performance of our website and the performance of our appliance website as well. That kind of feeds a lot into our product process. As I say, we have tons of ideas, but we don’t have very many people to throw at them, so we have to be pretty judicious about which ideas we decide to pursue and which ones we can set off to one side.
Usually it’s driven by three basic principles we kind of developed ourselves over the last four or five years of creating products. One is it has to work for almost any website. We don’t try to build products for a very specific niche. We try and make sure that it’s applicable to many different types of websites as possible. The second guideline is that it makes life easier for the website owner. And third is that it creates a better experience for the website visitor as well. Typically, what we do is we take an idea and once we’ve kind of summarized it in a clear way, we throw it against those three guidelines and say, “Okay, does it pass the sniff test?” And if it does, then we go about trying to validate it with feedback from users of our current products or just go out and find out who it’s useful to.
A lot of that is actually my job. I go out, and I just talk to influencers or I talk to website owners and just try to figure out what is actually helpful. That, funny enough, is exactly what led me to you, Joe. You popped up on my list of good people to talk to and I was just like, “Well let me see what this guy thinks of this product,” and then we got started talking.
Joe: Yeah, that’s exactly right. Along those lines for the validation process: you’ve tested it against your three tenants and maybe it’s approved going forward, but from what I’m hearing you don’t start coding, you don’t start designing yet. What specifically do you guys do in order to validate the product idea?
Jason: It really depends on the product to be honest. We start a few different ways, but I guess, and this is a process that we’ve been learning ourselves over a few cycles of product development and marketing and modernization and all that kind of good stuff. Where we’re at right now is, we start really as simple as possible. We’ve made the mistake in the past of shutting the door and closing the blinds and turning off all of our phones and just working for four months to create a product that we launch and then it turns out that it’s off the mark and we have to run pretty quickly to catch up and make a product that people actually want. As it works right now, what we do is we start almost with just pencil and paper sketches, we storyboard it out. It’s almost like how a movie comes together really. You don’t just shoot a movie without an idea of what it is that you want to do.
So we start with just back of the napkin sketches trying to describe it in different ways until we hear what we call “signal,” which means we put it out there and people really respond to it, they really share it, they’re enthusiastic about it. That gives us a signal. It’s almost like pinging a radar and when we get a ping back we’re like, “Oh! What’s that signal? Let’s go figure out what’s going on over there and see if there’s a viable product in there.” It’s basically just talking to people a lot of the time. When I describe a product to people, do they latch on to it immediately and say, “Oh my God I need that right now please build that,” or do they say, “Eh, you know, what else do you got?” Or, “Oh that’s like a different product,” and then they send us a link to a product that already exists or that could possibly compete with it or something like that. So yeah, it’s a lot of conversation.
Joe: I find that very interesting. I feel like with some of the different things that I work on it’s always a good sign when you’re excited about something and the people you’re showing it to, your prospective customers, are excited about it as well. It’s interesting to hear that you guys have found that.
Jason: Yeah. We try to be honest with ourselves as well. We ask ourselves, “Is this something that we really, really want to work on? Would we find this fun? Is it something that we would be excited to work on every single day?” Because building a product is a big commitment so it’s not just something where you can just phone it in every day or just do it for money. We’re not really about that. So there’s always that kind of important gut test as well. Is this something that we think we would be proud to work on?
Joe: I like that. As far as that research process, without giving away your secret formula, can you explain just a little bit more, kind of the nitty gritty practical side of step one where you might send out some surveys. Step two, you might actually talk to people on the phone for 30-45 minutes. I’m just curious to know how those conversations transpire and even how you decide who to talk to.
Jason: Let’s see. We do surveys every once in awhile and we try to keep them as short as possible honestly. The last survey I sent out was maybe three questions and a box for your email address if you wanted me to follow up on the responses. So I see who the most engaged people are, who leaves me their email address and who actually wants to talk to me a bit more. I follow up with them directly, usually via an email, sometimes with a phone conversation. It really depends. We use Qualaroo quite a lot, which is a great little app you just deploy on your site and it allows you to put up one or two question pop-up surveys. Then we keep an eye on the responses that come back in, and if we see a pattern emerge from multiple responses, we follow up on that. There’s kind of an unspoken rule of thumb that we follow, which is kind of the rule of fives or the rule of tens. Basically, it goes something along the lines of: If you hear three people say the same thing, say, in a survey, there’s probably another 30 behind them, i.e. 5 x or 10 x people behind them who are not responding but have the same sentiment or think the same thing. So 30 people who are thinking the same thing is a pretty significant number which means there’s a signal there. That’s one of the ways we try and derive signal.
I also do a lot of Skype interviews with users of our product as well and that’s really helpful for figuring out what features are really important and which one’s are not. When you talk to enough people, you can see a common thread across all of those responses and so the nitty gritty really is just what tools can you use to scale that and make efficient use of your time. I’m one person and we have I think four people dedicated to our products department. Scaling and efficiency is top of mind around here just because if it doesn’t scale or we can’t cover enough ground among the four of us it probably isn’t a useful devotion of our time.
Joe: What about the next step? After you’ve done some research, gathered some information, what’s the next step in the process?
Jason: Typically we start building prototypes, and then we will find people who are enthusiastic about the idea but will also give us good feedback and we’ll reach out to them and say, “Hey, would you like to participate in this private beta?” And we’ll build a prototype, keep it super rough, and just get feedback. It’s a constant learning process. As I said, this is not always the way that we’ve done it. This is kind of the place that we’ve arrived at given five years worth of spectacular failures and mad successes as well.
We try to involve our users at almost every point in the conception of a product. Just to make sure that we’re checking ourselves and building something that’s useful to other people as well as solving a problem that we have internally.
After the prototype, what we’ll do is we’ll typically just string it up with a bunch of analytics so that we can just see what is going on in there. There’s always a divide between self-reported data from users versus observed data. A lot of the time those two things can conflict. Someone will say, “Yes I would use this eight times a week.” Then you watch their activity and you see that they logged in like twice in the last month. That’s not to say that users are lying. It’s just sometimes you need to watch and observe in order to really get objective feedback. Other times, it’s great to just hear people say things that you’ve stuffed internally and haven’t prompted them with. That’s such a great feeling when someone says, “Ah, I would really love if your product would do x-y-z,” and we’ve already got people working on that feature because it means that we’re probably on the right track.
Joe: So where do you go from there? Is there usually a further iteration? How does the process work compared to knowing when it’s ready to be released to the public at large?
Jason: Yeah, let me think. We have a kind of a philosophy around here of betterment. I mentioned betterment already. There’s actually a process behind achieving betterment. It’s almost like a formula. A simple and compelling solution is the best way to remove friction and that equals betterment. What we do is we try to make our product as simple as possible and as easy to understand as possible. We also try to make it compelling as well so that it’s something that people enjoy using. It’s not just, “Oh I have to use this thing,” but they actually look forward to using it. Then we also look at points of friction. Friction can really come in a lot of different forms and it goes pretty deep, I guess as a philosophy, but at a very high level, it’s just stuff that people get hung up on. If a button is labeled in an unintelligible way or if something is taking too long to load or it’s too complicated for people to get through, that’s friction.
We try to remove that friction so that we have something that’s simple to use. It’s compelling, and it’s so easy that people don’t get hung up on it or get caught on things. Imagine strolling down a corridor and you’ve got all these obstacles in the way. We try and remove those obstacles so it’s a pleasant stroll down the corridor as opposed to an obstacle course.
That’s where our user expertise comes in and that’s how we methodically go about creating betterment. Even just working on one piece of that formula every single day creates betterment because if it’s simpler, that’s a better experience. If it’s more compelling, that’s also a better experience and if you remove friction that’s a better experience. We try and apply that formula as much as possible for our day-to-day.
Joe: I like what you mentioned about creating something that’s simple and compelling. I feel like that’s something that a lot of people set out to do. You might realize that a simple solution is the best, but I feel like that’s a lot easier in theory than it is in application.
Jason: That might be one of the most difficult things. Yeah. Especially in software. It’s so easy for software to become bloated.
Jason: You really have to fight for simplicity. And to keep something simple can also be incredibly complex. It would be incredible if all analytics tools, for example, boiled down what they provide into just what you need to know or interpreted it for you and told you, “Hey, this is what you need to do because we observed x-y-and z changes in your site.” As we’ve been finding, it’s incredibly difficult to build something that just deals with a complex topic like web analytics and performance down into a bite-sized sentence that tells you exactly what you need to do. That’s kind of why we’re trying to attack that as, “Yes it’s complex, but we think it will yield a lot of betterment for a lot of people when it comes to the engagement of their site.”
Joe: How do you, based on your experience, toe the line between adding more features or adding fewer features, and deciding which one of those is going to result in the most betterment? Is that a constant struggle or do you have process that you consider?
Jason: It’s always a balance, honestly. As I say, we don’t have a ton of resources to throw at just building feature after feature or app after app. So we really allow our users to kind of give us the signal. If something keeps on being mentioned … I do a lot of the product support for filament … that’s what keeps us clued in to what it is that people want. There are a couple of interesting things about being in support. When people request features, sometimes that’s not necessarily the feature that they’re asking for or that they need. It’s just that you need to present the feature in a different way or the language needs to be tuned to make sure that the people are seeing that you offer what they’re looking for.
Also, I have to balance those requests with what it takes to actually implement the feature as well, and that’s always a tricky balancing act because what seems like a very simple thing to people might actually be incredibly complex. I guess a recent example would be, we have a sharing bar app named Flare. We were getting constant requests for a mobile solution for it because it just wasn’t effective on a mobile screen. So we initially thought, “Oh yeah that will be easy. That’s fine. We can totally hop on that and deliver a kind of betterment just by addressing the needs.” Once we dove into it, we realized there’s a good reason why very few sharing bars have a decent responsive interface for sharing on mobile and tablets, it’s because it’s incredibly hard. You’ve got so many different screen sizes to account for.
Plus you also have to account for portrait and landscape mode, lots of different browsers, and a lot of sites don’t even cater to responsive. It was a tough obstacle course to navigate. Ultimately, we saw enough benefit in it to throw the resources needed at it, and we came out with a decent solution which is the Flare Pro.
Joe: Very interesting. I have one more question for the product development side, and then I can leave it open ended for you to add anything you feel like that’s been missed. Then I’ve got a few more broad questions that I’d like to ask you about. So the final question for product development and what we’ve been talking about so far: what’s your experience with knowing which features to build out like you mentioned. For example, you might have one person ask for feature X and you might have 30 people ask for feature Y. You realize, “Okay, there’s a pattern here and more people are asking for Y.” I’m curious to know what that process is like and also curious how you make sure you’re not doing a knee jerk reaction every time someone requests something where you start running around thinking you should build it. Do you have any advice about that?
Jason: Yeah, definitely. As I said, it’s always down to a kind of signal. We always are trying to see what is it that most people are requesting and what could we do that would result in the most betterment.
Signal isn’t necessarily a marching order. In the case that you described where you have three people asking for one thing and thirty people asking for another, it doesn’t automatically mean that the thirty people will get exactly what they want. That’s just a stronger signal, but at the same time we have to sit down as a team and discuss. Ultimately, it’s all driven by what makes the web overall better. We could have 30 people asking us for a pop-up builder, but we won’t build a pop-up builder because pop ups suck. It’s not necessarily just a numbers game in terms of what people are asking for. It’s what makes the web better because at the end of the day, the user experience of the web is what we’re trying to improve. So we use these kind of signals for us as guide posts, but they’re not necessarily marching orders.
Joe: Very good, that’s exactly what I wanted to know.
Jason: It’s funny because it would be easy for us to build what most people want, but at the end of the day, we probably wouldn’t enjoy working on those products and we probably wouldn’t be proud of them either. It’s almost like Steve Jobs said, which was that, “The customer doesn’t necessarily always know what they want, but it’s our job to give them what they need.” That’s the mark of any good consultant, any good doctor or service provider. Just really caring about what it is that people need as opposed to just what they’re asking for. Like I said, when they vocalize what they want it’s useful information for us to take into account.
Joe: I really like how you guys have specific benchmarks and tenants that you follow and different things have to fit inside of those, and that helps to guide you for making important decisions.
Jason: Yes. It’s a fascinating process and it’s something that we learn from as well. This didn’t quickly merge fully formed; this whole kind of take on product launches. We learned over the course of several different product launches and releases and maintaining them that this is what works for us.
Joe: I love how this all has come out of your experience not not just some kind of theory that was dreamed up. Okay so, I’ll leave it open ended. If there’s anything else you feel that that we missed from the product development side, feel free to weigh in on that. Another question I have for you is not specifically about digital products but product development at large. I’m curious to know if there are certain companies that you view as being really good at product development who are in this space and even companies not in the digital product space who you admire.
Jason: Let’s see. This is the kind of question that you always wish that someone would ask you in an interview, and then I don’t have anything lined up off the top of my head. I guess in terms of people we admire … Yes, we’re designers so Apple; yes Apple, definitely features very close to the top of that list. We look at people outside of just web design, honestly, I guess. We’re huge fans of Oculus Rift, what those guys are innovating is pretty spectacular. We just got our own Oculus Rift in the office which we’re going to plug in to various devices to play around with. In terms of design, the Medium team is really high up there as well. I just love what they’ve done with blogging and the experience there.
Joe: Awesome, that’s a great list.
Jason: Basically we love it when people try to improve things permanently by design. So, yeah, anytime that you see stuff around like that, chances are we’re all over it and sharing it internally. I’ll stuff emails with links in it, and we’re constantly finding interesting new stuff. Those are some of the people that kind of stick out in my mind.
Joe: That’s great. Well those are all the questions I have. I don’t know if there’s more you want to say or if you’ve covered everything.
Jason: Nope. I really appreciate the opportunity. I hope maybe it’s useful for everybody, for all of your listeners. If they’re interested in checking out Flare, it’s filament.io/flare. I would love to hear what you guys think about it, honestly. As I said, the feedback is always taken into account. So any comments that you guys have would be very much appreciated.
Joe: Sounds good. I really appreciate you taking the time. I enjoyed the interview and learned a lot. I’m looking forward to publishing it and getting it out there for our readers to take a look at it.
Jason: All right. Well thanks a lot, Joe.
Joe: All right. Talk to you later.
Joe Putnam is the blog editor at iSpionage. You can get his latest PPC and CRO advice on Twitter at @josephputnam.