Sign up for our newsletter
Technology / Data

Analysing Big Data: Q&A with Recommind

Bob Tennant, Recommind

How is business holding up during the recession?
The short answer is great. Over the last few years we’ve basically doubled revenue every year and we’re going to do a little better than that this year. The reason is that particularly in hard times if you have solutions that help generate revenue and reduce costs and in particular help to lower risk, those are solutions that will do well.

We’ve also had a significant effort over a long period of time to evolve the platform in a way that can be helpful for the new field of analytics of large volumes of unstructured data and we have a few announcements about that over the next few months.

So that is the Big Data movement?
Living in the Valley we get inundated with it and there is a lot of smoke and mirrors at the moment. It’s very easy to become marketing buzzword compliant; it’s not very easy to deliver on something. If you look at what people would label Big Data it’s a huge variety of stuff. It basically has anything to do with large volumes of data. Analysts will talk about the four Vs: volume, variety, velocity and value of data.

White papers from our partners

What it really is all about is as volumes of data have increased it becomes more and more difficult to manage. But that’s also an opportunity – there are kinds of things you can do if you have very large volumes of data that you cannot do if you don’t. That really comes along with analysis. If you have very large volumes you can start to do statistical analysis.

But that’s moving towards the business intelligence side of things, isn’t it?
In a sense. The way I think about traditional BI is that, for the most part, it’s counting and drawing pictures. There isn’t a whole lot of analysis there.

So as the volumes of data have gone up there are opportunities to exploit it. A lot of people have taken their existing software and rebadged it and others that have tried to build highly-scalable platforms to handle big volumes of data, but in many cases it’s been for the same kinds of problems they used to solve, just at greater volume. That seems to me a difficult way to build a business.

So for example, until recently, Oracle didn’t have something they would call a Big Data platform, they would figure out a way to build a highly-scalable relational database system. There are at least 10 start-ups in Silicon Valley doing that but that’s not necessarily a great avenue.

On the other hand there are great opportunities in two areas: one is unstructured information and the other is in analysis, whether that’s structured, unstructured or a combination of both. So when we think about our business there is the problem and the opportunity. We have focused on unstructured information and I don’t think there are many companies doing a good job of that and there are even fewer doing a good job on analysis of that unstructured information.

We’ve had an initiative underway for about a year to effectively form a new kind of database which allows us to tie together structured and unstructured information to form the basis of analysis.

So what are some of the uses of this?
Fraud detection comes up a lot. It has been around for a long period of time and it is relatively easy to do when you’re dealing with structured data. But being able to identify key data within large volumes of unstructured information is new. That is where you will see a lot of innovation; the marriage of unstructured and structured and its analysis.

What’s the key to getting that right?
The key is being able to identify the key data; you have to understand what you’re pulling out. That is not going to be as easy as a lot of people think and you will see a lot of announcements from vendors that will not be particularly helpful.

The reason is that in an unstructured data context the data is application-specific. So while it is relatively easy to do things like identify people, places and so on in text it’s not very easy to identify whether or not a particular person or entity is related to a particular problem. If you can do a good job of identifying what the key information is you can do a good job of analysing it.

One of the criticisms of these systems is that they are not cheap or easy to run.
I think both of those are true. I think the first is only a problem though if there is no value being derived from the system. The problem with generic platforms is that it is difficult to take a sophisticated set of technology, drop on the average user’s desk and tell them to use it. They don’t know how to take advantage of the situation.

So the approach we taken is to focus on solving problems in a specific manner. We’ve figured out how to apply the technology to a particular problem, as opposed to giving them a toolkit letting them build it themselves.

One of the other hot topics in IT at the moment is social media, which is creating huge amounts of unstructured data. Is that forming a big part of your business yet?
It’s an opportunity for us at the moment. There are issues there with people utilising those technologies in ways that are not always appropriate but it’s also an opportunity from an analysis side. If you can tie together information that is being published on social sites with information in the CRM platform or the marketing platform you can do more interesting analysis.

So what about the consumerisation of IT and the impact the likes of Google, Facebook and Twitter are having on the way people use information?
When I think of consumerisation I think of it as making it easy for people to solve their problems. Those companies you mention solve different problems. So while they are easy to interact with Facebook does not do what Google does which does not do what Twitter does. People find things easier if the application is suited to purpose. I don’t try to shoehorn something Google provides me with into what Facebook does; it’s too unintuitive.

I guess Autonomy is your main competition. What did you make of their acquisition by HP?
I think it was the right idea [from HP] with probably the wrong asset. They didn’t try to do anything too dissimilar to what IBM did so successfully. The problem is if you sit down and try to work out a way to screw it up execution-wise, you could not come up with a better way. On the same day you announce you’re going to pay that price for Autonomy you also say you’re going to get rid of the biggest part of your business [the PC division]. In the long run it may prove to be not a bad thing.

The challenge with Autonomy is that it was viewed as a high-growth company, but it wasn’t, organically. They were buying a lot of stuff and that increased revenue.

There has been a lot of M&A activity in the information management space recently. Is Recommind next on the list?
We’ve received phone calls periodically throughout our history. The challenge for us is that we are growing rapidly that it makes it unlikely we’ll be acquired. You are more likely to be acquired if growth slows because the multiples then come down.

Such as with Autonomy?
The multiple that HP paid for Autonomy was huge. It’s the kind of multiple you’d pay for a high growth, organic company. Autonomy wasn’t that. Frankly, HP probably overpaid. The way Autonomy maintained its top line growth rate was to buy assets. Like clockwork every year they would but a bigger asset, and that worked great for them. But there’s a reason the market responded to that deal the way it did.

Is there much demand for your technology on mobile devices such as iPads?
It depends on the use. You wouldn’t dream of doing an eDiscovery review on a phone. On the other hand if you are in the field a lot and are sending and receiving emails you want to be able to interact in that way. So there is a lot of demand for our information management applications on mobile devices. There is a lot less demand in other areas.

Are you guys looking at going public at all?
We are looking at going public, probably not too long from now. The macro-environment is one of the things we’ll be keeping an eye on. The company is in a position to go public if we want to. There’s a lot of uncertainty in the world at the moment and that might not be so good for public markets but it’s great for Recommind.

Does the reaction to the likes of Facebook’s IPO not worry you?
There has been a significant decline in the stocks of what are not really IT companies, basically consumer-oriented software-delivered services like Facebook and Pandora. They haven’t done well at all as a group. On the other hand if you look at enterprise software companies like Splunk and DemandWare, they’ve done extremely well, they are up significantly up as a group.

I personally think there is a bubble in the consumer tech space. Everybody hears about Facebook or uses it so they think it must be growing like crazy but if you look at the multiple they are paying on top of Facebook’s revenue you’d have to be Apple to justify it.
This article is from the CBROnline archive: some formatting and images may not be present.