Datamine extracts the facts from the figures

So the data processing aspect of your business is a done deal, eh? Information about customers is pouring in by the bucket-load and accumulating on hard disks out the back somewhere.

But that doesn't necessarily mean it's doing any good. Collecting data is one thing - analysing it and making decisions informed by fact is another matter altogether, and not necessarily one that a general IT department can help with.

Specialised expertise is called for: Enter Datamine, which Paul O'Connor started some 10 years ago. The company has since grown to about 19 employees, most of them engaged in data analysis for a range of mostly large clients here and overseas.

"My background is a little bit of statistics, a little bit of marketing, and some business studies and computer science as well," Mr O'Connor says.

"Data mining was very much an emerging field in those days, so we had to struggle for customers at first. But now we have a good mix. We mix those disciplines effectively: we understand business problems, we know how to handle big databases and analyse them for business relevance, and on the stats side we can develop models and effectively analyse information."

Sally Carey ("We don't like job titles at Datamine; everyone knows what they're supposed to do" - but she seems to manage the business) says she sees Datamine's role as helping people to make better business decisions by using data they've got available to them.

"It's a little hard to understand if you're not in our business but an example is the Department of Corrections," she says, picking a case study that has in the past not always made good decisions and is, presumably , now trying to lift its game. We've recently helped them forecast their prison populations and therefore their staffing requirements.

"What happened before was the predictions they had were wildly inaccurate, which made their future business decisions very difficult," says Mr O'Connor (who also doesn't have a formal title, but who seems to have a considerable knowledge of computational statistics).

Their predictions weren't accurate even for a year ahead, which had serious cost implications."

"The sorts of organisations we're working for are government departments and blue-chip companies," says Ms Carey. "We provide advice, consulting and education - a lot of the time businesses don't know what data they may already have and how they can use it."

These days Datamine finds that its clients mostly seek it out, without much advertising being required, although the company does send speakers to relevant conferences when it can.

"We rather naively went to speak at conferences in Melbourne, Sydney and Brisbane this year," Ms Carey says, "because we do mean to expand into the Australian market.

"We had such a flood of enquiries that we had put most of them off - we were only doing it to spread awareness. Now we're managing it through an intermediary - it's almost as if we've got Australian business despite ourselves.

"We also have clients in Asia, starting from work we've been doing for a New Zealand company with an office in Singapore. That has provided us with contacts and work across the region - Korea, Hong Kong, Japan, China, Taiwan and India."

"A lot of business decisions are made on hunches, and we're trying to make sure there's some facts behind them," Mr O'Connor says. "Most organisations have pretty good data, but getting from raw information to something that can indicate a better course of action is the tricky part.

"What makes us different is that we're willing to deal with literally billions of pieces of information and make sense of it - we're beyond Excel spreadsheets and into large, grunty databases.

"One customer base we're looking at for an Indian client has around 180 million customers. It's a different scale of work. When they want to understand who they should be addressing with their products and who they should be marketing to, they ask us to discover who in that database is most likely to take up each of their various products and address them specifically. You can't do an unspecific mail-out to that many customers."

"Convincing clients was difficult back then, and in some ways hasn't got much easier. There's still an education piece.

"For example, one of our customers a few months back was looking at dropping one of their brands and was basing this decision on some research done five years ago. We were able to show them that their two brands were actually delivering quite well but if they dropped one of them their customers would not necessarily migrate to their other brand.

"We introduced some factual rigour into their decision process."