The last few years have seen a flurry of activity in the business intelligence (BI) markets. Firstly, the old-style approach of reporting against data in a relatively static manner was put under pressure by users demanding something a lot more flexible, where they did not need to approach the IT department every time they needed a new report creating.
Then, the “big” vendors decided that they needed BI tools that they could control, rather than embedding an OEM copy of a standard engine (generally Crystal Reports) into their systems. Therefore, we saw IBM acquire Cognos and SPSS, Oracle Hyperion and SAP Business Objects. Even some smaller vendors got in on the act – relative newcomer Spotfire went to Tibco, the middleware vendor.
This flurry of activity presented issues to many users – if they were running a heterogeneous environment, how rich would the functionality of one vendor’s BI engine be against another (competitive) vendor’s systems? Would bringing Cognos into an Oracle environment be a foot in the door for IBM to try and move the organisation over to DB2?
The space in the market for independent BI providers soon began to fill up – vendors such as Information Builders, Panorama Software and Tableau Software have all been making noise to let users know that there are alternatives available to whatever they may be using to date. The largest remaining independent vendor, SAS Institute continues to dominate as the largest player by far for BI from a specialist vendor.
Cloud-based computing is giving greater reach to many of these vendors – for example, Spotfire has had a hosted version of its software for many years, and Panorama (through a partnership with Google) offers its capabilities as a BI software as a service (SaaS) offering.
But the biggest problem for many remains the cost of BI. If a business’ main asset is its information, and therefore its intellectual property, there is need for access to BI to be widespread across its employee base, so that everyone can add value to the data at their disposal. By providing BI tools only to those deemed as being of a high enough employment level for the cost to be acceptable, BI ceases to deliver what it promises – and becomes a niche tool.
But, open source could provide the answer here. There is a broad open source BI project, known as BIRT, the business intelligence and reporting tools project. Actuate has the best known distribution of this, but others provide alternatives under open source banners that may offer greater overall functionality.
The main players here are JasperSoft and Pentaho, both of which have “community editions” of their software that are free to download and implement, with support being community based. Their commercial editions are enterprise subscription licences based on number of cpus, so broad adoption can be made without the need to deal with per-user issues.
Others not using open source but having low cost approaches (and often a free usage version or at least free evaluation version) include QlikTech and Panopticon. When it comes to choosing the right BI direction for your organisation, I recommend the following as basic pointers:
- Is a single vendor solution important to you? If so, look to the tools from the incumbent (e.g. IBM, Oracle, SAP), but negotiate hard around existing contracts. On the whole, this will tend to restrict you to a smaller number of active BI seats, with the capability to distribute the reports from the analysis.
- Do you want an independent BI system that can work across a more heterogeneous system? Look to the remaining independents or semi-independents such as SAS Institute and Tibco Spotfire. These may still be expensive at a per seat level, but will often come with tools to manage deep domain expertise areas, such as oil and gas, pharmaceutical and finance data analysis.
- Are you able to provide your own systems support, and are you in a position to enable a suitable number of support staff to build up their own knowledge of the chosen system? If so, then community editions of BI tools may meet your requirements. This would provide the most scalable BI system – but you need to watch out for any lack of functionality that may be in some community editions.
- Do you want a fully supported system, but also want broad adoption capabilities? Then look for supported versions of open source tools, or for commercial systems that do not charge on a per user basis.
For many, it will come down to a hybrid solution – but certain key areas must be borne in mind:
- Everyone must be analysing the same information. Different BI tools looking at different information silos do not provide business intelligence – it gives chaos.
- One person’s great BI tool is another’s incapability to see the results. Bar and pie charts may be pretty easy for everyone to understand, but not all data is well suited to bar or pie charts. Heat maps, spider charts, geo-data representations, bubble charts, you name it may all have meanings to certain people, but do not assume that any one way of viewing data will suit everyone. Look for tools that enable the user to change the visualisation method to one that suits them best. If multiple BI tools are chosen to give different viewing capabilities – see the point above re information silos.
- “Live” reports have pros and cons – web-based output which allows users to see the analysis of the live data is great – unless what was required was a snapshot in time. Make sure that the differences are understood, and the right reports can be run against the right data.
Finally, buyers must understand that there is a cost to everything. Free, open source software (FOSS) is not free – even when you take the community version. It is free of licence, but has implementation and support costs. Subscription software has on-going, predictable costs, and commercial, off-the-shelf software (COSS) may have upfront licence costs along with on-going maintenance charges. The final choice comes down to other issues – as detailed above – and applying suitable business wisdom in the search for meaningful intelligence.