³ÉÈË¿ìÊÖ

Google has offered businesses the chance to use its servers to crunch huge amounts of their raw information.

The firm's BigQuery service is designed to help organisations identify and analyse trends from their datasets.

Google said small businesses struggled to access such tools in the past.

Experts said that the service had the potential to disrupt the analytics industry, but questioned whether firms would upload their data to the search giant.

Google said BigQuery could process billions of rows of data in seconds. It said results would be offered as downloadable files or could be viewed via its cloud storage division.

The firm carried out a small trial of the service earlier in the year, and has now invited other organisations to sign up to a wider preview offered free of charge for a limited period.

A outlined how the service could be used.

"Imagine a large online retailer that wants to provide better product recommendations by analysing website usage and purchase patterns from millions of website visits," product manager Ju-kay Kwek wrote.

He also gave an example of a car manufacturer using BigQuery to explore how a marketing campaign performed by studying "billions of multimedia impressions".

The big firms already involved in the analytics industry include SAS, IBM, Oracle, SAP and Microsoft. There are also many niche player offering specialised types of insight.

A recent study by companiesandmarkets.com suggested the sector was set to grow by an average 6.8% a year between 2010 and 2014.

Analysts described Google's intervention as "aggressive", but said Google might struggle to become a big player.

"Generating analytical results of high significance for business decisions has not been easy historically," said Christian Lagerling at the technology analysts GP Bullhound.

"It typically requires significant PhD level engineering hours to be able to fully comprehend and apply the results.

"Security may also be an issue, not only how Google handles the data but also in terms of authority to use an 'outsourced' tool like this."