ITProPortal got up close and personal with Dai Clegg of Acunu Analytics, a firm that specialises in squeezing useful information out of big data. One of its biggest clients is Hailo, the taxi app that has graduated from London's Tech City and gone on to conquer much of the globe. We were keen to find out more about Acunu, real-time analytics and the behind-the-scenes work required to power a successful company in today's non-stop world of business.
What do you do in your role as VP of product marketing at Acunu?
Two things. One is traditional marketing - which is finding the people who might be interested in what we've got, whittling out the ones that aren't and then creating meetings and connections between people who have maybe never heard of Acunu before. Then our sales team can help them understand what our technology can do for them. The other part of the job is communicating to the world what Acunu is about. For example, the space we're in is what we call real-time analytics. Analytics of extreme data. For Hailo, that means that its Acunu systems can tell it where the hotspots in London right now are. They can see that visually.
Also other simple things, like how long processes take after people hail a cab from Hailo. When you've got an app on your phone, if it isn't slick, you'll never use it again, particularly when you know there are at least eight cab apps in London. If Hailo wasn't working for you, you'd just move on. So it's very important for the company to know if its app is working well. Do customers get a swift response? Do the cabbies get what they want? We measure their KPIs in real time. And that's pretty much consistent for all Acunu customers. There's always something people need to know about their businesses, and they need to know it immediately. We provide a way to capture all that data and process and visualise it in a very short amount of time. Traditional business intelligence tools send captured data into an operational system, where it's extracted, transformed, loaded into a data warehouse and so on. The trouble is, it might be two or three days old by the time you get there. For lots of applications in business that's perfectly okay - if I'm comparing this quarter to last quarter, I don't need my data to be perfectly fresh - but with things like mobile apps and click-stream analysis it's different.
How closely do you work together with hailo?
I feel that we took Hailo's advice: to build as little product as necessary to get your customer, and then build what your customer wants. It's not quite as simple as that, but that was very much our attitude. "Well, we think this is a great idea - what do you think Hailo?" They took it, they bashed it and they found the good bits of it. We tried to be very responsive.
How central to Hailo's success do you think Acunu has been?
The real key to Hailo's success was that it had a great idea, a very strong management team and it executed well. We were a part of that execution. Hailo basically said, "We have this problem and those guys can fix it," so it plugged us in and we fixed it. I wouldn't say Hailo would have failed if Acunu hadn't existed - it would have either built [the tool] itself or found somebody else. But it might have taken longer or cost more and might have eaten into Hailo's time to market and margins.
How valuable is data analytics to businesses? Is it the same for organisations of all sizes?
I think it's massively important. Analytics is the dashboard on the car. Without that it's completely operational - you can change gears, you can hit the accelerator and the brake, you can drive - but you do not know what condition the car is in. You don't know if you're about to run out of fuel. Without Sat Nav you don't know when the next turn is. The analytics is all the information you need in order to make good decisions when you're driving a business. You can start off without it in a sense but if you don't have a vision, a management team and funding, you'll never get there. So a lot of people start without too much in the way of analytics, but that doesn't last for long.
Not all analytics has to be what I've described as real-time. There's a whole spectrum of business intelligence, from monthly reports right to who's clicking on a website right now to monitoring whether the behaviour of our customers is being affected by that marketing offer we launched online this morning. Being able to make better decisions quicker is massively important. and I think that's becoming very common now across all industries. If you were to ask people about big data a year ago, they'd have said, "That's Hadoop." Hadoop was the first wave of big data which made it cheaper to break down big information, therefore people could take on more more quickly. If you ask people now, real-time is the area that's growing. 12 months ago it was all about Hadoop and there was a little bit of interest in real-time. Now, the interest in Hadoop and batch analytics is still there but the interest in real-time is growing.
The intriguing thing about this is that whereas Hadoop in many cases was taking existing use cases from data warehouses and implementing them better, bigger, faster and more cheaply, real-time analytics are often providing insights people just didn't have before. They didn't know these things about their business. Before Hailo used Acunu, it didn't know the answers to those questions. They can now make better decisions quicker because they simply know more about their business. The reason they weren't doing many of these real-time use cases before is because they were just too difficult and too expensive. It's only new technologies - things like Cassandra - that have made it possible to address those kinds of use cases. We suddenly realised we were meeting loads of customers using Cassandra to try and build something like Acunu Analytics themselves and we'd help this customer and that customer do it. We suddenly realised that if we just built a product version of it, they'd all get it done quicker and it would lead to a better margin for us. We didn't set out to build a great product and then try and find someone to try and sell it to. We started building and then people pulled us into it. Real-time analytics is not about me telling you that it's important, but our customers telling us that it's important.
At GigaOM Structure:Europe, Tim Moreton said that outsourcing big data analytics can be problematic, since the issue of ownership can become blurry. Do you agree?
What we've seen - particularly with financial organisations but also with others - is that they're not willing to give their data away to a third party, for that third party to then pore over and do the analytics for them. Analytics is all about having a huge amount of data and doing something clever with it. So if you want an outsourcer to do your analytics, you have to give them access to your data. A lot of organisations are not quite ready to hand over masses of data to cloud platforms. I'm not the best person to judge, but my instinct is that this is slowly eroding. But it is certainly still true today, and that's what Tim was talking about. The problem is this ownership. That's why Acunu Analytics does not run software as a service in the cloud. You licence it and you install it. You can install it on a cloud server if you want to - many of our customers install on their own servers on-premises. It's that issue of analytics giving away the data.
Acunu recently launched Acunu Analytics 5.0 for Cassandra NoSQL databases. Can you tell me more about the 'self service' aspects of it?
At the Cassandra EU Summit we announced self-service BI. If you work in the marketing department and you want to understand the marketing data the IT department gives to you, you'd expect to be able to run some sort of cool tool that allows you to visualise that data and build charts, reports and summaries to help you understand it. You don't expect to have to say, "Hello Mr IT, can you write a report for me?" You'd be told to join the queue which could be three months long, after which you might find out that you haven't actually got back what you wanted.
If you take the world of NoSQL, like Cassandra, the tooling levels are low. What we did was to make it possible for the same person to say, "I want to see my Cassandra data, and I'll write the dashboards, reports, summaries, charts and maps." At the launch, Acunu deliberately made me, the marketing guy, perform the demonstration to prove that you don't have to be too technical to do it! I took a bunch of tweets about Cassandra and Acunu and I built a dashboard that monitored the tweeters' whereabouts, placed them on a map and colour-coded the different countries. It took me about a minute and a half to build that app, when all I had to begin with was a file full of tweets. Acunu Analytics didn't even know what a tweet was until I gave it a file and it worked it out. Until now, if you wanted to do that with data warehouse data, you had to drag and drop. You can use our tool now, and that's a big step forward. It means that just two people can do the job, so it's faster, and the business people get what they want because they do it themselves.
What does the future hold for Acunu's relationship with Hailo?
Hailo is clearly a global success and a fantastic poster-child for London and UK startups. We'd like to ride on Hailo's coat-tails to a certain extent. If we can let them blaze a trail and build the technology that doesn't slow them down, then that will be a success for us and we hope that will contribute to Hailo's success too.