Stalking the landscape

Huge volumes of transactional information and a formidable competitor in Microsoft are both a threat and an opportunity for business intelligence vendors. How will they cope?
1 November 2007

How will business intelligence tools help companies, especially telcos, with terabytes of transactional information? Will Microsoft`s recent moves on the BI front hurt or help the existing vendors? Present at a Brainstorm  round table discussion to give their answers to these and other questions were: Caron Mooney of IS Partners; Marc Scheepbouwer, MD of Intellient; Adrian van der Merwe, MD of 8th Man Consulting; Richard Greyling, managing consultant at Knowledge Integration Dynamics; Carlo Gunter, COO of Institute; Nitesh Vallabh, director at PBT; Dillon Gray, channel manager of Business Objects; John Olsson, sales and marketing director at Ability Solutions; Filip Vanden Houte of MCI Consultants; Greg Bogiages, director of Cortell; Estelle de Beer of BI Practice at Sybase; Erwin Bisschops, senior solutions architect at Harvey Jones; and Christo Bredenkamp of Synergy Computing.

Brainstorm: How will BI tools help telco providers move towards transactional-based revenue rather than the current access models?

Richard Greyling: We have major telco customers and most of them have felt that to analyse their transactional information at this stage of their projects is too much. They`d rather summarise it and not overanalyse it at that level. One of them has the largest data warehouse in the southern hemisphere – 320 terabytes and growing daily. Analysing that is very optimistic. They will have to get to that level eventually to be competitive, but at this stage in the South African market, it isn`t such a risk. Luckily the guys in Europe are doing the groundwork so we`ll be able to learn from them.

You need a different animal for hardcore BI solutions.
John Olsson, Ability Solutions
Nitesh Vallabh: In certain aspects, telcos are using transactional information, particularly in product development, where they need to test how a product will perform in the market. They take a sample data set and put it through a set of parameters to see what usage would be like for a certain product.

Estelle de Beer: BI tools are capable of analysing large volumes of data – it`s the back-end systems that need to be capable of producing result sets on those large volumes where there`s often a bottleneck. The BI tools can give you the transactional information, but if you can`t store 14 terabytes and allow them to interrogate that, then it becomes a back-end infrastructural problem rather than a business problem.

Dillon Gray: There are two distinct markets in the telco environment in this country: landline and cellphone. If we look at the mobile environment, we have so many applications: SMS, WAP, marketing and so on. So there`s a need for the carriers to start reinventing themselves to bring in that information and try to use it as a tool. But, as has been said, competition on that side doesn`t exist in this country. Is there really a need for them to do it? Right now, I don`t think so. Our landline operator is under a lot of pressure, both internally and externally. Internally it has legacy people who have a set way of thinking about what BI should be; externally, they have new competitors coming in. From our perspective, the landline operator needs BI to reinvent itself and BI will also help the competitors. But the telcos now want BI more to do their billing rather than look for transactional revenue.

Marc Scheepbouwer: We have a cellular provider as a customer and they use a BI architecture to analyse quality around calls. It`s tied to a culture in that business to deliver quality service. They take all the call stats, drive it through a data warehouse and an analytical layer, and present the results to their mast engineers so they can pre-emptively manage their masts. So, BI is being used in some areas. Although the technologies are capable, there are scalability issues. It is challenging, not only at the transactional level, but also for the interpretation of the data.

Christo Bredenkamp: One of the areas in which we see BI making a difference to mobile companies is in their call centres. There we`re dealing with operational data and it is possible to do live analysis of it, such as [the number and frequency] of dropped calls.

Erwin Bisschops: I was involved with a large project with a telco overseas where they used BI tools to predict churn, which is quite an important KPI. It was a BI environment set up in a closed loop, with their transactional environment linked to a marketing module. The data in the warehouse would be analysed by marketing people and the business logic could be applied in that environment.

It produced a list of people who would be walking away within the next three months. That list could be fed directly into the marketing campaign module and the responses would flow back to the warehouse to check its effectiveness.

Greyling: The business model for telcos in the UK and Europe is substantially different to that in South Africa. Here you pay per minute; there it`s a fixed cost. Here, there`s little value in moving from one [operator] to another, whereas over there they`re competing on service.

Brainstorm:  Microsoft`s moves in the BI space: good or bad for the industry?

Caron Mooney: Microsoft started in BI around 1996, so it currently has a fairly mature third-generation platform. It is starting to eat a bit of our lunch – by buying our suppliers, for example. It is definitely causing a pervasiveness of BI to the extent that other vendors are being forced to drop their prices or deliver better value or take a different angle. The company is definitely democratising BI.

The downside is that because it is pervasive, everyone with three months` experience calls themselves a BI developer. The market doesn`t understand the difference between solid BI experience and a developer who knows how to spell it. It`s the same thing with a lot of technologies: they were specialised and are now a lot broader, so there will be a lot of bad implementations because of that. But there will still continue to be good ones.

Bisschops: Services have exploded since Microsoft acquired ProClarity last year. It is interesting to see that Microsoft has good tools to offer, but that`s not the reason why people buy it: 44 percent buy Microsoft because of the price.

Greyling: Microsoft`s biggest selling point in this market is price; you can`t compete. If the company is implementing on a Microsoft database then what tools do you sell? It`s very difficult to convince a customer to part with more money. The only time you can justify it is at big enterprise customers. Microsoft hasn`t quite got there yet. The other side of it is that the most widely used BI tool in the world is Excel. In that way, Microsoft is actually competing with itself.

John Olsson: Obviously, Microsoft`s model is to commoditise everything. At the risk of being controversial, the original bad image of BI comes from traditional ERP vendors – post-Y2K and post dotcom – where they were looking for new revenue streams and to cross-sell and up-sell. The one thing people have learned as the market has matured is that you need a different animal for hardcore BI skills. It shouldn`t be the same consultant that deploys an accounting system or an ERP system.

De Beer: Microsoft has made our jobs easier by educating the customers as to what they can get out of BI, but it`s also given BI a bad name because of poor implementations. A customer said to me recently that ‘you get what you pay for`. So they are still willing to pay for an enterprise BI solution because they see it as a specialised skill. I don`t know how that will change in the future, but right now, enterprise customers have the perception that you need specialised tools and skills.

Carlo Gunter: In our market, more SMEs are starting to open up to BI and making it more affordable. Not that it hasn`t been affordable for a small company, but there was a perception that it was only for big corporates. It`s really a case of SMEs getting exposed to BI; once they experience what it can do, they`re open to moving to a more specialist solution.

Greyling: I have a sizeable manufacturing customer that is using Microsoft for the data warehouse and other tools for the front-end. Based on the costs and the licence fees, they are seriously considering moving to Microsoft once it`s enterprise-ready. Telkom is embarking on the Microsoft route and adoption is going to happen in this country. I think the other vendors are very worried about it.

Vallabh: Why is it that there always has to be one tool, while being Microsoft all over? Co-existence is the order of the day when there are pros and cons to every product. I have this debate with my customers: why are we talking about the tools? We all know the methodology, so can this tool do what you want to achieve, yes or no? If it can, keep it; if not, throw it out. If I want to do high-end analytics, I`m not going to use Excel. A manager who needs to make four or five decisions a month doesn`t need to do regression analysis to make them.

Mooney: It`s part of the purchasing policies that IT places on vendors and it`s also Microsoft`s licensing policies. A lot of people buy it because it`s cheaper. IT likes policies. The bigger the company, the more streamlined the policies: one tool, one support base, one development base.

Vallabh: What concerns me is that the same principles used to buy a building or furniture are not the same principles for buying BI solutions, especially the high-end stuff. It`s like looking for diamonds in a sandbox: you may or may not find them. The issue is that as you get better at finding them, you need to adjust how you do it. If software is not decreasing your costs and increasing your revenue, then throw it away.

Gray: With Microsoft`s interest in the market, it has forced the vendors to look at their market differently in even simple scenarios such as activity-based costing, one of the crucial aspects of any telco environment. Vendors are looking at the core environment and saying that there are a lot of players, so let`s look outside and see where else we can bring tools to market.

Greg Bogiages: We never try to compete head-on with Microsoft on price. In many instances, where we come across Microsoft tools, customers have no idea that it`s actually BI. Microsoft has always limited itself by going into a company through the IT department. They talk to the technology guys, saying, ‘We have the whole suite across the board with cubes and whatever other tools`. But when you speak to the finance manager, he has no idea what they`re talking about.

Bredenkamp: The specialist vendors have the added advantage of experience – how BI should be positioned, sold and implemented, so Microsoft has quite a bit to learn – but everyone should still be sitting up and taking notice.

Olsson: Microsoft will continue to commoditise the products and ship them as a total BI solution. You can see within their own business solutions where they ship entry-level cubes to gain a foothold in the market and that will threaten some of the specialists in the medium term. There`s certainly a huge opportunity for everyone around the table here because we deliver services far beyond what you`ll get with a standard Microsoft product.

More SMEs are starting to open up to BI.
Carlo Gunter, Institute
Filip Vanden Houte: The thing about the improvement of Microsoft`s BI tools is that it allows skilled people to get more done in less time. When Microsoft`s tools improve, it puts pressure on other vendors to improve features and functionality. That`s a good thing.

Bisschops: BI isn`t liked because it`s BI. It`s liked because the operational environment doesn`t do things properly. One of the main challenges for BI in this country is that it needs to be used properly. I`ve seen a site where a BI tool has been abused as a reporting tool because SAP R/3 can`t do pretty prints.

Reinvention around the corner?

As consolidation in the BI market stabilises, two fascinating revolutions promise fireworks in the segment next year. At the high end, telcos with their staggering volumes of transactional data will try to apply BI tools and techniques to garner some of that transactional revenue necessary for their long-term survival. And at the lower end, armies of SMEs will start seeing it as a commodity. If it results in the availability of more BI skills locally and a more competitive market, bring it on.