The business intelligence (BI) market has evolved over the past three decades. In the 1980s, BI was really just about reports, which were given to everyone, focusing on what happened in the past.
Nick Felton, head of business analytics, Advanced Business Solutions
Conor Shaw, vice-president general business and partner ecosystem, SAP
Phil Davies, channel and alliances director, EMEA & APAC, Pentaho
John Sands, product sales enablement manager, Qlik
Edward Smith, technical director, Intuitive Business Intelligence
Mike Hallett, alliances and channels director, Oracle
In the 1990s, vendors recognised that there was demand from customers to go beyond reporting and drill into that data to work out why things happened and the way they worked – and that is when a lot of desktop ?reporting tools emerged, and Excel became the premier BI tool in the marketplace.
The past couple of years have seen the addition of visual discovery tools. In the 2000s, vendors realised that just giving ‘power users’ tools was not delivering much value, so they ushered in the era of dashboards to address the larger audience of causal users, so more people could use data. It was also not just about looking in the rear-view mirror but looking at how things were working at the time and how they were tracking to plan.
Then, in 2010, the industry switched from looking to the past and present to using all the data it had been collecting and applying it to mathematical models to predict what could happen.
In the past three decades, every 10 years the market has shifted focus, just like clockwork, to another part of the BI market. Vendors have flipped from focusing on the traditional power user – between 5% and 20% of the traditional audience for BI tools – to the casual user, the executive, the manager and the front-line workers who use information to do their jobs but are not hired to crunch data. Reporting is the world of casual users and the solutions are more IT-driven, allowing people to see their key indicators at a high level with the option of drilling down.
Q.The market has a history of change, so do you see that continuing right now and what sort of changes are you detecting in the market
Conor Shaw: The machine is doing what was previously done by the human. In the 1980s, analysis was being done by the analyst or manager looking at a data dump. But then technology improvements came along that meant they only had to look at certain things, so then it became monitoring. Automation of monitoring of processes frees up users to use the tools to explore and analyse the data in ways that they would previously not have had time for, such as predictive capabilities.
John Sands: It is about making predictive analytics available to the user and making sure that not just the casual user, and not just the power user, but everyone right at the very edges of the business can get that analysis. It’s about making sure that these complex algorithms that run in the background and do all of this predictive analysis are being made available to all users. They don’t have to know how, but they do need to know the answers. Data scientists are a rare breed and not every customer has one, but the algorithms they create can be available for all organisations.
Shaw: It is about user trust because at each stage of the market development, the user trusts the technology to do analyses that they previously did themselves. If you take prediction, at the moment predictive analytics says that if you do nothing, this will happen, but you can start to say, “I would like these outcomes”, and then find out what factors are the most influential in arriving at the desired outcome and hence understand what you need to change to reach that outcome. Users are looking for this level of predictive capability and even using the automation available through “machine learning”.
Phil Davies: If you look at financial services and trading, the machines are already making all the decisions. They are just using BI and the growing volumes and types of data available to make smarter decisions automatically based on established rules. The next step is about execution.
Q.Where do you think the channel partners and customers are in terms of ?the sale and adoption of the latest BI technologies
Nick Felton: We have such a vast range that some partners are still in the 1990s, but others are at the cutting edge.
Sands: A lot of customers still want the reporting and want that PowerPoint presentation on their desk at the end of every month.
Felton: There are variations because it depends on which part of the business you are dealing with.
Sands: The systems integrator channel is much more advanced, like Atos and Capgemini, and they are looking at the future and are very much there and are selling it, particularly into the public sector.
Shaw: I would look at our traditional partners – they have a lot of legacy systems and installed base to deal with. However, some new emerging partners are embracing the new, innovative technologies. In this industry, if you don’t change, you die. If our partners are not getting up to speed on Hana, mobile and predictive analytics and if they are not executing on the cloud strategy, then they need to think about building those capabilities and we will have a conversation with them to help the change be accelerated.
Mike Hallett: Some people have failed to retire what they were using in previous decades, so
they have ended up with an estate of legacy stuff and a lot of silos. So, some want to be
forward-thinkers looking at predictive analytics, but many still have old reporting procedures
running in the background. What we are seeing is a trend for partners to help people to modernise
and integrate the silos.
Q. Casual users and power users make up the audience for BI and require different types of tool. Casual users like to use dashboards and reports, and power users are using discovery tools to find insights. The two biggest segments are the visual discovery vendors and traditional online analytical processing (OLAP), which is used by casual users. Where will the visual discovery market go? Will it keep growing and where will it sit in the market?
Shaw: There is a shift and big data is enabling operational analytics, competitive maintenance and that sort of new thinking. Customers want to start to put that into products out of the box. The CFO or CMO running their business will start to put that back into the organisation. With the Hana platform, we can process data very quickly, so you can make informed decisions with the data that is produced in a time that is useful. Big data analytics is going to explode. The casual user will want to know and you can’t leave it with the data scientist. They have to provide the infrastructure for the manager or user to utilise and interrogate. The casual user will want to use their iPad or mobile to find out if the weather is going to change tomorrow so they need to order more ice-cream, or whatever they need. The way the reports are delivered is down to personal preference.
Felton: At the moment, we have customers that are looking for forecasting and the planning elements. The BI element, the dashboards, have largely become a commoditised piece in terms of software. Most companies we talk to have many different versions, but the key differentiator is around the forecasting and plans they can execute to have an immediate impact on the business. Also, people in back-office operations want to look at sales, inventory and stock levels and to get a view of what that does from a P&L perspective. Many people have not reached this stage. They are still looking at the P&L with Excel and have some of the dashboards, but haven’t got that bit to be able to execute.
Davies: Visual discovery will continue to grow as analytics platforms support more types of data. However, it will always be constrained by earlier stages in the data “production line” where data cleansing, storage, preparation and blending take place – and this can get very complex. The “eye candy” in reports and dashboards always helps to sell BI, but these are no good if the data behind them can’t be trusted. Channel partners have a huge opportunity to help customers set up and manage the whole data management production line. When we asked people what types of applications they were using for predictive analytics in 2006, it was marketing intelligence about cross-sell, upsell and campaign management. Today it’s budgeting, and that’s the dominant application.
Hallett: No one knows what will happen in the future, so we need plans to say what could happen and what should happen. If you ask most business people where their targets and business objectives are modelled, many today would still turn to something in a spreadsheet. So, we have partners helping the business adopt better collaborative planning and budgeting applications.
Shaw: It is about adding the context around big data. Some people have gone in with traditional discovery tools but had to rip them up because they have not been delivering. Big data is great and many people are talking about it, although not many know what it is. But without any context, knowing that you sold 1,000 blue t-shirts on a Tuesday is not much good.
Davies: I asked one of my partners what his definition of big data was. He called it “speed data”. It goes back to that idea of giving the casual user the power to make decisions quicker. Speed data was a very apt description.
This was first published in September 2014