Consumerisation and Application Development
In this, the second, part of MIS Asia's interview with Donald Farmer, VP for Product Management at business intelligence (BI) technology provider QlikTech, the impact of consumerisation on IS organisations in general and BI tools development and deployment in particular are discussed. The expurgated transcript below.
Talk about consumerisation.
Consumerisation is important for two reasons. The first is just the impact of devices. The estimates in our own business are that within a year or two years, about 25 percent of all business intelligence will be mobile. And that is being driven not by IT departments but by decision makers in business. When you get something like an iPhone for your birthday, you'll probably then want to use it in business. I can't remember the exact number...but was something like 60 percent of all iPads are bought by individuals and 80 percent of them are used for business.
Those people are buying their own devices. Then you have the "bring your own device" phenomenon. Now, the interesting thing is that it is also happening in software. Gartner has identified a trend of people buying their own software. They have also said that we are leading this, by the way. Business users are choosing their software and, using a phrase that Gartner uses, they are doing it "with or without IT's permission." In other words, they're going to their IT departments and saying to them: "Give me the software, let me use it." There's the example of the guy in the airport starting to download software in the airport and solving his problems before his plane landed. He didn't phone IT to ask, "Can I install this?" He just installed it himself. So we're moving from "bring your own device" to "bring your own software." And that's a huge change in consumerisation.
The other part of this is consumer behaviour. It's different and that is because consumers have learnt. Consumers learn a certain type of behaviour. They search and they browse, rather than write queries. They expect to browse an application rather than go through exact commands. Think of a search engine. Today in Google, I just enter "Search" and I get information back. If I don't like what I get, I change and refine my terms and search again. I might add a few words. I'll say "Hotels in Singapore", then "Business Hotels in Singapore", and then "Cheap business hotels in Singapore". I'm constantly refining my process. I'm refining it all the time. I'm not constructing the perfect query.
A few years ago, if I was using Alta Vista or I-Cost, I would actually be very careful to say the "hotels AND Singapore AND business AND Marina Bay". I would construct a query. It was like writing a query in a query language. None of these matters now, because we have learnt to "browse". Consumers are used to browsing and they want to browse. They don't want to have to learn the query and construct a query, they want to just browse. That means that you ought to be able to give them the data quickly. The great thing about Google is that it does two things amazingly well. It gives you lot of information and it give you lots of information very, very quickly so that you can interact with it. We have to do that. And that is very different from the approach of a SAS or Microsoft.
The other aspect of consumerisation or consumer behaviour is the emphasis on apps, and being able to have small and specific applications to solve your problems. Those you can just go download and work with very, very quickly. That is a very different approach to software. Previously with software, there was an acquisition. Even if there was a person buying software, he would have to choose the application carefully. It would probably cost a lot of money to buy an application, and take time to learn to use, especially so if you wanted to use all its features. Nowadays, you go to the App Store to download something-it could cost you anything from nothing, it's free, to just a few dollars-and you expect to be able to use it within minutes. Imagine if you downloaded an app and then had to go for a training course to use it. An app that needs a training course-that doesn't make sense.
That's what we make sure to avoid. We don't have training courses for our apps. We have training courses for the developers, yes. But the business user who just browses the QlikView app, he doesn't need training to use it.
Real-time analytics. I'm sure you've heard that expression. What does it mean to you?
It means a lot of questions because what is real-time? If I have any customer that talks to me about real-time, it raises a lot of questions from me. What do you mean by real-time? I don't know what you mean but every customer is going to have a different answer. You have to make a distinction between two types of analytics or intelligence when you talk about real-time. There is streaming analytics-analytics over data that is constantly pouring out of some system and the importance is to analyse the stream-or there is very fast analytics, where every time you query or update the analytics, you're actually updating or looking at an entire data centre.
The reason why those two things are different is first, I have to ask the customers, "What happens when your network cable gets pulled out by accident or your network goes down or the Internet connection is not working? What do you want to happen? What happens about the data that is no longer streamed?"
If the answer is, "Well I will have to be able to pick it up again, and be able to go back in time and recalculate that stream and catch up with all that streaming data." Then that's one type of real-time analytics, because I have to be able to handle the fact that the streaming breaks.
If on the other hand they say. "It's okay because the next time I do a query it will just catch up with that data, then that's a different kind of real-time analytics." Then my response would be, "So is it a stream or is it fast queries? If it's fast queries, we do that very, very well and people can use QlikView in those environments quite successfully. But if it's streaming, that's a specialised area which is probably better handled by CEP [Complex Event Processing] and applications or by some of the highly specialised streaming analytics products that are out there." We have a partner that does a lot of work in financial services and they are present here in Singapore. They have some customers here. They integrate with us quite successfully. They provide streaming analytics, and so we can refer them to our customers if the need arises. We are quite happy with real-time analytics questions. However, our initial response is often more questions than answers.