Lean and mean the order of the day for BI
By Mia Andric
Many IT markets work in cycles. For example, the centralized infrastructure of the mainframe was replaced with the broad de-centralizing effect of the PC, and the pendulum has swung back to centralized again as cloud services on central hosts take over. In the Business Intelligence (BI) market, focused best of breed products were the norm, followed by a period where many were acquired by the larger vendors and incorporated into a full stack. Recently, best of breed BI tools are back on the up. Premie Naicker, MD of Yellowfin South Africa says she’s seen a really big shift from the full stack platform-based BI back to best of breed.
Business users pushing best of breed BI approach
“The pushback has come from business users,” says Naicker. “Businesses are throwing up their hands and saying this doesn’t work for us. To the business user, they seem a bit let down by the BI – by the tools and the results – because it’s not innovating. The existing approach to BI in many companies, together with complexity of it, has meant that what happens is that business users are the ones that work around IT and get something that works for them. This ranges from using tools that aren’t part of the enterprise platform, to only using a tiny element of what’s available.”
Traditional BI megavendors struggle with legacy code
Naicker says the reason that full stack BI isn’t innovating as much as the smaller players isn’t because the vendors wouldn’t like to, but because these so-called ‘megavendors’ have an enormous amount of legacy code that cannot be rewritten – or thrown out.
“Some large stacks even contain COBOL code that was written in the 1970s,” she says. “But the code simply cannot cope with the tremendous innovation over the last five years. Innovation that can be typified by the introduction of the iPad and consequent influx of Mobile BI. So the megavendors have a big problem: the legacy technology on which their platforms were built isn’t compatible with many demands of contemporary BI development. That, and they also have to grapple with their large installed base. In contrast, BI and analytics vendor Yellowfin started off in 2003 as a server-based application written in Java by BI professionals. It still is – Yellowfin hasn’t had to revolutionize the technology on which it’s based to cope with the sweeping changes that are taking place in workplace BI usage.”
Changing BI user demographics
Naicker says the other driver favoring the more agile player is the wider range of customers using BI products. While BI was once seen as another of the IT department’s projects, today’s business users are actively evaluating BI tools in this age of Big Data.
Hype around Big Data ’revolution’ blunting BI spend
However, Big Data is quickly approaching the Peak of Inflated Expectations as measured on Gartner’s Hype Cycle. Companies of all sizes think they need it and, as a result, many vendors are offering solutions that can crunch billions, or even trillions, of data points quickly so that businesses can unlock value. But the reality is that most organizations don’t want to spend too much on Big Data because it’s risky – they’re not sure what they’ll get back.
“At every Big Data conference I’ve been to, I’ve spoken to very few attendees who have got real value from it,” says Naicker. “What the Big Data hype has done is created a far more complex marketplace for analyzing information: people are not spending money on Business Intelligence, but they’re not spending money on Big Data either. Everyone is paralyzed and waiting to see what happens. I think the reason is that Big Data solutions can deliver little nuggets of insight for any organization but little nuggets are just that: little, and you can’t run a business on them.”
She adds that marketing hype around Big Data has resulted in confusion between the terms “Big Data” “BI” and “analytics”. And that, with the hype around Big Data’s potentially revolutionary nature, organisations are hanging back, waiting to work out how they can use Big Data in a revolutionary way.
“The reality is that it’s most useful when applied to existing business challenges, in order to reach a result with a greater level of granularity,” says Naicker.
For successful business analytics initiatives, look to data at hand
Naicker comments that the vast majority of organizations have all the information they need to understand their customers.
“Banks and retailers have transactional systems and customer service departments already. All they have to do is bring them all together – to create a unified, trustworthy and analyzable view of their data – to understand what their customers are doing and how to generate leads.”
Combine location-based and regular data for context and deeper insight
These days, one of the ways in which companies can better understand their clients is through the location-based applications that abound on mobile devices. However, Naicker points out that this is one of the least used but most powerful aspects of company data: “Customers have a location. So do suppliers. So do your markets. More than 70% of your data has a location component. Traditional BI tools typically just offer tables, grids and charts, but this only tells part of the story.
With map visualizations, you can quickly relate BI data to locations that are meaningful to your business and detect geographic trends, such as customer clusters or outliers – there’s no Big Data crunching required because you already have all this information.”