facebook-pixel
  • Big Data & Technology
  • Marilyn de Villiers
  • MAY 23, 2019

Seven must-haves for business intelligence success

There are seven key factors that can mean the difference between an analytics project succeeding, or adding to the high statistic of big data project failures; a failure rate Gartner puts at 85%.

The first step towards business intelligence success is for data to be regarded as an asset.

The first step towards business intelligence success is for data to be regarded as an asset.

This is according to Karl Dinkelmann, director of data enablement, business intelligence (BI) and analytics at AccTech Systems, who told delegates at ITWeb Business Intelligence and Analytics Summit 2019 this week that all indications point to the fact that any company embarking on an analytics project has an 8.5 out of 10 chance of failing.

"Analytics projects fail not because the solution doesn't work, but because the business fails to realise value from its investment, or the technology is not used at all. The cost of this failure is enormous," he said.

The first step towards having analytics take its rightful place in the organisation is for data to be regarded as an asset, on par with every other asset owned by the business. And like other assets, data has to be properly maintained and managed, and this has to be done by the people in the business who actually use the data.

"It is important that the data is managed properly because if the wrong data is entered into an analytics program, all that will happen is that you will get the wrong answer faster, and this will destroy trust in analytics," he said.

In order to get the best from an investment in analytics and ensure it delivers real value to the organisation, Dinkelmann outlined seven practices that should be followed.

Assess

Conduct a data maturity assessment, which includes listing all the data/complexity areas seen as important.

This could include issue tracking and monitoring; developing and maintaining a data catalogue; determining who is responsible for data stewardship; data modelling; data quality; data storage and so on.

Build a picture of how it should look and keep it on your radar as issues to work towards at all times.

Obtain guidance

When choosing the analytics products needed for an analytics project, do not rely on product punters who are trying to sell their product without taking into account what the business really requires.

It is also important to look closely at the management requirements of the proposed solution. Establish whether you will be able to maintain it yourself, or whether you may be locked in to the supplier and what it will take for the organisation to become sufficiently mature to be able to deal with all the associated management issues.

Set a strategy

This is important as it will enable the company to focus on data maturity issues and what it will take to get there.

Acquire resources

A fairly broad range of skills will be required to ensure the success of a BI project. In addition to BI skills, a business will need people who have experience in data architecture; master data management, business intelligence, and data warehousing; and data science or advanced analytics.

Develop data domain architecture

Very often, much of the development effort can be spent on cleansing data. But in addition to data quality, look at data modelling, data cataloguing, the report register and physical architecture, the target operating model and the master data.

Involve the business

Establish lines of responsibility and define who is responsible for the data. Set up a data governance body that includes data stewards on the one hand, and those with data management competency on the other, and ensures they work together.

Manage the change

Ensure firm data principles are in place. Without these principles, there will be false starts, frustration, gradual change which is all but impossible to manage, anxiety, confusion, and unsupported change.

The Harvard Innovation Lab

Made in Boston @

The Harvard Innovation Lab

350

Matching Providers

Matching providers 2
comments powered by Disqus.