In San Francisco (CA), we meet Co-Founder of Crittercism, Andrew Levy. Andrew talks about his story how he came up with the idea and founded Crittercism, how the current business model works, as well as he provides some advice for young entrepreneurs.


Martin: Hi, today we are in San Francisco with Crittercism. Andrew, who are you and what do you do?

Andrew: So I am the co-founder of the company. I lead strategy and basically handle all outward facing activities and I also work with the team on product strategy and roadmap.

Martin: How did you come up with this business idea?

Andrew: Basically, I have an engineering background. A Long time ago I was doing data warehousing at HP so big data type engineering work. I left there when I got into Y Combinator, a startup incubator based out here in the Valley for a different company. There we were working on a project, it’s a data pipeline that analyzes tweets to be able to target advertising. We are quite passionate about ad tech space so we ended up shutting down that company. It was my first start-up after college I was working on but I met up with a coworker at HP and we started building mobile apps together. Our apps, they had a bunch of performance problems, we had one star reviews, we had people complaining. We just found out it was very difficult to get a handle on what that end user experience was like. We ended up building some tools to help ourselves and we spoke at bunch of mobile meetups and realized that this was a widespread problem.

The apps we have used ourselves were not that actionable. I am sure you’ve read that before, they say, “This app sucks. It’s not working.” Ok, it doesn’t help me recreate the problem and even if you could get in touch with the consumer, they just aren’t knowledgeable to help you understand and recreate exactly what happened and what happened under the hood. That’s what the company is for, we solve problems, scratch your own edge is the phrase. So we are building apps back in 2010 and we ended up incorporating this company in January of 2011 so actually comes up on our 5 anniversary now.

Martin: Did you have some kind of initial customers before you incorporated the company and this made you see that there might be some business potential? Or did you first incorporate it and then went looking for customers?

Andrew: It’s a great question because after Y Combinator we got a little bit of money from there but we were bootstrapping and my fiancée at that time was buying groceries which was great, she was being very supportive, but we wanted to make sure that the next company that I build, that there was something there, that there was a real problem, a ‘hair on fire’ problem that needed to be solved. After we’re building those mobile apps we almost didn’t start this company. But we felt like there was something there and so what we did was we set up a sign up list, like a beta list, put up a landing page and that was just a way for us to gauge demand . And in addition to that we did get software out in the hands of a few startups, a few other developers that we knew in the community. We just saw that there was enough demand and we actually took that and applied to a different incubator called Angel Pad which is founded by a bunch of ex-googlers who are now based in New York. We got into that program. So we had the social proof and social validation through that. Plus we had that sign up list of potential clients and so we said, “We have to do this. We are on to something.”


Martin: Let’s talk about the business model of Crittercism. What is actually the value proposition that you are delivering to your customers?

Andrew: It’s actually fairly straightforward which I think when you work on something, can you explain it to your grandma – that’s kind of a test, right? Basically everyone has used a mobile app that hasn’t worked correctly, maybe it was slow or laggy, maybe tried to buy something and it didn’t work. So what Crittercism does is that it attract that user journey while they are using a mobile app. It sees where you struggle and we collect diagnostic data under the hood, to help companies find and fix those issues that customers are running into.

I guess a bit of a deeper explanation of what we do is to really provide actual intelligence by taking operational and behavioral data. So behavioral data is what buttons the user is tapping on, what screens they are on. And then the operational data that we overlay on top of it is code defects – things like crashes and errors, networking issues, not only carrier performance but we track all the cloud services and APIs that a mobile app is interacting with. And we tie it back of course with code defects because they can be related. We look at important transactions or workflows in the app so if you are trying to log in or update your profile or check out, we take that and we track all these critical events and help you understand its taking too long to complete an action and if there is some failure that occurred along the way.

Martin: When you are talking of user behavior tracking is it that the company or your customer is actually defining the workflow or is It like generic like maybe Google Analytics maybe just add the Crittercism snippet and that’s it?

Andrew: It’s similar to google analytics. We have an SDK piece of software that you embed into an app. We automate a lot of these events that we collect and a lot of the performance issues that a customer may run into. We do an optional logging so you can send us additional data and also we let you define what those important workflows are. Our clients know better than us where does logging start and end so they can define that and we will automatically track all these critical events, user behavior and operational data along the way.

Martin: Are you having an API so your customer has access to all the people data as well?

Andrew: Yes, we have an API for all virtual data and our best clients will hook us up into their operation centers. The product managers will have real time view especially around releases. The develper team will have a separate dashboard, seeing what the code defects are like and how well they are doing.

Martin: Are there any specific customer segments that you are having or is it just everybody who has a mobile app?

Andrew: So we operate across verticals, across industries. We operate at a tremendous scale. We process over a hundred billion application launches every month. This is a across three of the five top global media companies, two of the top three hotel chains in the world, two of the top three credit card processors and so it’s not every single vertical. We do have major presence in retail. I’d anyone that’s building a mission critical or revenue critical mobile application falls under our purview. And it’s in a variety of use cases, and it is not necessarily who thinks of the consumer apps when you download apps at the app store, that is the majority of our business but a lot of companies are also publishing internal mobile applications to their employeesto improve a business process.

So an example would be— we have a very large retailer who has applied apps to the employees in the warehouses so they can track inventory.There is another one that we work with, and you have seen point of sale devices being done by smart phones or tablets, but they actually will even do something like take a smartphone and hook it up to a printer and reprint price tags when people return items to the store. So we are seeing more and more of that, all these business processes are being moved onto smartphones and tablets because of the efficiency gains that they get from it.

Martin How is your revenue model working?

Andrew: So we charge based on how many users an app has. Specifically, monthly active users. If you install and never use the app we won’t charge the company that’s supporting the app for that user. But it makes sense because we collect more data the more customers that client ends up having.

Martin: And how did you define the price point or is it really like customer specific?

Andrew: It took us a while to get to something that worked because there wasn’t a great analogy in the beginning. It wasn’t like we can look at what was happening on the web world, because mobile is just different beast. And so we just experimented and over time found a price point that worked.

Martin: Ok, cool. Are there any options to extending this kind of model because if you are first on the mobile and you have all the behavioral data so to speak where can identify wether an app crashed or some system information and etc. what you are currently doing. Are there any other applications that can build on top of that or actually you are already doing?

Andrew: Sure. You are right, we collect a lot of rich data. In fact, we started publishing some of this in the forms of reports in The Health of the Industry, as iOS and Android is doing. In fact, we’ve got, we have some live benchmarks there. So we will continue to invest in the data intelligence aspect of it.

But we have our customers using us in a variety of ways, ways we couldn’t even imagine when we first built the company. At the end of the day it’s all about customer experience and companies want to make sure that no matter what channel their users are going through to access their service that they have that amazing experience. And of course there is revenue involved to make sure that everything goes smoothly.

So an example that would be obviously around connected devices and IOT, we are in some major watch applications, as an example we have just seen Apple revamp its TV offering. So these are platforms that we’ve started to work on and we plan to support all of them because they actually look and feel and smell very similar to smartphone and tablets and there really you are dealing with embedded software. And the important thing to keep in mind is when working on embedded software is customer experience is very tightly coupled with performance because everything is combined into one. You don’t have that client – server separation of the web. And companies just struggle to keep up because the ecosystem is moving so quickly, there is not enough engineering talent out there, they can’t even hire fast enough. They have caught up somewhere, on the traditional mobile side but now you have this explosion of brand new devices coming out and it’s another challenge. Our software helps them tackle those challenges so their customers are successful .

Martin: You are covering one and one hundred billion app loads a month. Can you give us insights on the main reasons why an app is crashing?

Andrew: Sure. And when you say crash, you know people use this term for variety of things. When you look at an app crashing, it could be just close but what happens if you try to buy something and it’s sitting there and spinning and you say screw that and hit the home button some people might call it a crash even though technically there is a word for it.

When it comes down to it, it’s actually related to why mobile is different. If you can answer that question, you can answer what causes apps to fail the most. When you look at it, you look at all the devices out there, you have heard a lot about device fragmentation, thousands and thousands Android devices. Even on iOS , you see they have a lot more devices than they used to. You see a lot of operating system pushes so we just had iOS9, Marshmallow is having a roll out right now. And each new iOS release can cause problems, each new device that comes out. You have carriers and a lot of people think carriers are dumb pipes but they can actually mess with the data, they can do package shaping we have seen one that have changed JPEGs to PNGs as an example. You have geopolitical issues like the firewall in China, you have major countries just shutting down internet access during conflicts. These are all things you just didn’t even have to imagine in the previous world. But of course the sensors available to these devices are new. And so we added all up, it is impossible to test every possible use case and even if you could that new Samsung device comes out and suddenly things start breaking again. Having that visibility, it is a vague answer but that’s the number one reason why applications fail. It’s just that there are peculiarities; whether it’s the device, the OS, the location that cause some of them to occur or some code to be executed that you just didn’t test for. That’s what ends up causing a lot of headaches especially for companies that that are not using our software.

Martin: Are you aware of the Safe Harbour Act? I think some weeks ago it was refused—

Andrew: Last week. Yes, The European High Court nullified it.

Martin: Your company is based in the US. Do you see any impact on the Safe Harbour on your business?

Andrew: So Europe has been important for us from day one. In fact, our first paying customer was in Europe which a lot of people don’t know. Even though we are 60 person company we have a European data center. We leverage Amazon’s Frankfurt Instance and that has data isolation so the data does not leave the EU. In fact, it doesn’t leave Germany, it’s based there. We were Safe Harbour certified before the European Court did that annulment, by the way we are paying close attention because they are going to give new guidelines to US companies, kind of like a Safe Harbour 2.0.

But data privacy and security has always been important to us, it’s always been a question that has been raised. Especially because we are embedded in people’s apps and so by default we don’t collect any personal identifiable information. In fact, we have turned a lot of capabilities off by default in our SDK and companies can chose to turn things on, depending on if they are comfortable with it or not. You have to deal with mobiles that can be used by children, especially children under 13 and there is a law against collecting data there.

So we have made sure to be compliant across the spectrum in the US and overseas as well. So I think other companies that maybe don’t have a European data center are probably pretty worried because when they sell to a company in Europe, that was one of the certification they would show to make them feel more trusted an feel better by using their product but for us, we have that capability available so there is less of an impact for us.

Martin: Just checking if I got it right, in the US you are not allowed to track the behavior of people of 13 years and below?

Andrew: There is a specific rule called Code of Compliance. There are specific rules around collecting anything that might be considered as person identifiable. We don’t collect that by default anyway but there is a gray area in terms of what you consider person identifiable and so we have made sure that we architected ourselves both in terms of our product but also in terms of the legal documents that we handle.


Martin: What type of advice could you give to first time entrepreneurs who would just come to you and say, “Andrew, I want to start a company but I am not sure what to do and how to do it correctly”?

Andrew: There are many pieces of advice I could give. I would say for one thing when we were trying to start a company before I think we were going about it in the wrong way and kind of sitting in a room brainstorming about, “Hey, what should we work on”. But when it really came down to the most successful one we were leaning on our own experiences in what pains we did encounter, what inefficiencies that we did see. It is hard to start a company when you are not intimately familiar, you don’t have the main knowledge. If you are a software developer and you want to create software for medical profession but you don’t have any medical background it’s tough, you need to find a co-founder that could fill that slot for you.

But beyond that, I think one thing that helped was as we started the company, one of the mantras that you hear quite frequently is ‘fail fast’. And even after we started this company we had what we thought was one product line but it was actually two and we begin quickly deprecating and getting rid of one of them because it wasn’t going as quickly. We wanted to really focus our time so I think it is important that we did that. Over time we have continued to make sure we stay focused and get rid of projects that are not on our core.

Another piece of advice that I’d give is that you are very scrappy at the beginning and you don’t have a lot of money and it really forces you to make decisions in the right way and to move quickly. So another piece of advice I’d give is as you raise money from investors and you start to grow the company is you keep that mentality. If you do raise some money don’t overspend, make sure you validate your sales model. Don’t go out and hire a hundred sales people before you have done that. The analogy might be some people say, “If you build it, they will come”. But what you should do is to let some clients come and then build a little bit more. Don’t burry your head in the sand and build something for a year and hope that you will be successful. You need to quickly iterate and that’s not just in software development but that’s also on your business model and go to market plans.

Martin: Andrew, thank you so much for your time. It was very nice and interesting.

Andrew: Thanks. It was nice meeting you.

Martin: And if you have an app and you want to really understand why it fails or where you can improve it on your workflow maybe you can just install and buy Crittercism.

Comments are closed.