Data Warehousing: House Of The Rising Sun

OK, so data warehousing isn?t the most accurately dubbed technology ever, but resellers that look beyond the silly name might just spot a serious business opportunity, says Guy Clapperton

There?s a certain fashion in vogue most of the time for deriding branding exercises. Someone spends a fortune on a new logo and the rest of the world says ?so what?? Someone else specifies a new product category and the rest of the world says ?huh??

So it has appeared in the case of data warehousing for some time. Given that it consists mostly of organising data in what could be described as an organised fashion, you can begin to understand the cynicism. Linguists hate it too, because of the mixed metaphors ? data mining gets introduced, as do datamarts. Whoever went mining in a warehouse, or set up a mart while they were at it? And why weren?t databases set up like that in the first place?

But in spite of all that, data warehousing represents a serious business opportunity to the switched-on reseller. As with all reasonably new technologies, though, making money is largely a matter of hitting on the right standard ? and getting the customer to part with their money in the first place.

The investment is considerable, the immediate return on investment (ROI) negligible. Nevertheless, IDC figures published recently point to a market worth $54 million in 1995, $90 million in 1996, followed by projections from the Meta Group that suggest an $8 billion market for 1998 ? to include services, software, hardware and every other attached sale, so it?s not a like-for-like comparison.

Probably the first issue a reseller needs to address, in view of this sort of growth, is not so much how big they need to be to cope with the market, but how substantial a customer needs to be before the technology is of any use to them. At first glance it appears that only the massive enterprise-wide installation is going to be able to take advantage of it, whereas the smaller user will be managing their own data.

Not so, say at least two suppliers. Mike Evans is managing director of Timeline, whose product is based on SQL Server, and offers a completely scalable solution ? don?t they all. But Timeline offers a single-user data warehouse to start off with. ?This might look like a tempting entry point for resellers wanting to test the water. It will recognise their ledger application and create the warehouse for them automatically.?

Although fighting shy of a single-user version, Cognos is convinced that businesses with turnovers as small as #1 million can find a use for the technology if their circumstances are right. Chris Coan, divisional director for channels, believes the essence of a successful data warehouse installation is finding the right application for it. ?One of our resellers uses our tools to help them bring their products to the market. They use it to manage their sales history and it is a benefit to their efficiency,? he says. Most of the applications to which data warehousing is put, though, are on a larger scale.

Clearly, a pitch at a market at this level is not a straight product sale. For many, the first step to overcome is persuading the customer that they need to invest the money to prepare their data in the first place, before anything like actual kit has started coming through the door. Price Waterhouse has been quoted as saying a standard 60 per cent of the cost of an implementation is accounted for by this preparation.

This is a view backed up by IBM. ?It?s daunting,? says David Topping, worldwide software marketing manager for data. ?You think it would be a good idea but you?ve got five terabytes of data, so where do you start??

This is where everyone begins to make ?start by buying my product? noises. IBM?s approach is reasonable enough, though. It tries to get its product, Visual Warehouse, into a target site to solve one particular business problem.

The theory is that a departmental project acts as a test bed for a corporate client, which then implements the solution throughout an organisation having seen short-term ROI. Long-term ROI can be impressive. The aforementioned IDC report suggests that an ROI of 401 per cent over three years, given a spend of $2.2 million, is no more than average.

So much for the customer benefits ? but the reseller still has to decide whether they are suited to this sort of sale. If Sybase, database and relational leave you cold, the chances are you won?t go much on data warehousing either.

Ian Macdonald, VP of European marketing for Planning Sciences, suggests the opportunities for selling Intel-based servers ? highly appropriate technology for handling the data in its raw form ? are unlikely to go away, while the chance to sell consultancy in terms of organising the data should not be neglected by the indirect channel.

Data Warehousing has its roots in relational database technology, which (according to one marketer with vested interests) was much better at data entry than it was at presenting the data afterwards. This is why data warehousing types rarely stoop to discussing the sale of storage products (although reports in the US indicate that because of the advanced state of their market, storage capacity is starting to take over as the chief cause of teeth-gnashing). Data is useless unless it is accessible. There are two business opportunities here: datamarts and tools.

There has been a lot of fuss made about datamarts, and SQL Group?s product director Pauline Dunne advocates these scaled-down information repositories as a useful entry-level for the smaller reseller trying to get a piece of the market. ?It takes all products a while to mature so that you can buy them out of a box, and data warehousing has been around long enough,? she says.

To which a possible answer is yes, but if your client is still an enterprise client then all of these datamarts have to be linked up again, usually by a big datamart. The key here is to understand the total needs of the client and, if the brief is to serve a department, to ensure that the buyer doing the spec knows what the rest of the company is doing.

The other means of ensuring that data can be extracted and organised, which must reside alongside the enterprise linkage, are the software tools. Topping puts it this way: ?A warehouse is only any good if there is someone making notes on what?s in it and where things are, down to which shelf a particular widget is stored on. It?s like having a warehouse without a stock list.? Tools and add-ons are a clear area in which resellers can add value to an existing implementation.

For future developments in warehousing, one need look no further than the internet, according to Macdonald. Planning Sciences? Gentia product is now fully Java-compliant and will work across intranets as well as more readily identifiable database technology.

This, he says, is an excellent means of implementing the product for companies such as credit checking agencies, which will want multidimensional information that is swiftly available.

The idea is that multidimensional information (information available in different categories, based on spending levels, geography and other criteria) will be available to them as a paid-for service.

Microsoft, of course, has an interest in the development. Alongside NCR and others it has set up the Microsoft Alliance for Data Warehousing, and has succeeded in developing an upgrade path from SQL Server to NCR?s Terradata environment. This alliance has yielded precious little so far but the Gates empire still appears to have its sights on data warehousing as a long-term prospect and a lot of other suppliers use SQL Server as their database of choice.

This is almost certainly an area to watch, although the chances of anything happening quickly must be as promising as James Goldsmith?s chances of being prime minister .

And by the way, it?s still a very silly name.