I talk to clients all the time about this topic. Advanced analytics programmes fit all 3 criteria. The stakes are really high – business life or death in many cases. Very few people are truly familiar with success in this field – often confusing it with traditional analytics. The potential for embarrassment is limitless – I’ve seen enough red-faced CEOs to know.

Here’s my top 5 guidelines when planning your data programme;

1.Leaders must develop a clear vision for the program

At Board level, you need to understand what the possibilities of advanced analytics really are beyond the hype of the media and vendors. It’s much more than traditional analytics but possibly less than where your imagination takes you initially.

Cutting through the noise is not so easy to do. Everyone talks a great game so talking to someone who’s been through this challenge successfully several times is your best bet. The trouble is that since there is such a high failure rate of these programmes, it can be difficult to find such people.

Once reality is understood, you need to create and then communicate a clear vision to give direction to your people and harness the enthusiasm of the leadership team.

2.Creating a productive analytics strategy

Your 3-5 years business strategy informs an effective analytics strategy. To be clear, advanced analytics is a set of sophisticated tools and methods to help you achieve your strategic business goals. Many businesses jump in without making the connection between the 2 or simply don’t have a business strategy that most senior people recognise. It’s going to be the first area of questioning I ask of any client.

Assuming you do, then the key to success is developing a big-picture, structured, long-term analytics strategy alongside your shorter-term initiatives that impact profits.

3.Develop multiple initial use cases

In my experience, the most successful data initiatives start when companies understand the entire decision chain of their business process and identify which decisions, if improved, would add the most value and have the greatest impact in the next 6-12 months.

Once identified, these handful of key decisions are analysed as use cases to understand the business value they could generate, prioritising the most valuable whilst checking feasibility (crucial). The feasibility checklist must include data availability and appropriateness, resources required vs available, disruption likely to be caused and finding someone enthusiastic to own the initiative.

Following these steps will give you the best chance of success, choosing the right project, creating a worthwhile impact on the bottom line and bringing the business with you on this initiative and the next.

4.Get yourself some analytics story tellers!

It’s absolutely critical to find those people who have the ability to really understand data from the business and tech perspective AND be able to tell a story. That last part is essential as they will not only translate thoughts and ideas between the 2 areas, but they need to be able to stimulate and enthuse both parties. They should then be actively involved in the project to keep the story evolving and people engaged.

They also need to feel supported by the data leadership team to have the confidence to believe in what they are ‘selling’.

This is a great job but a difficult skill to find. Look for the story tellers, the brilliant communicators in your business and train the skills where necessary. It really will make the difference between success and failure.

5.Define clear analytics roles

When creating a new data and analytics programme, early steps should include an open-minded assessment of the talents available across the business to deliver it. Some people are obvious already in the exact same or similar roles. However, some are hidden elsewhere in the organisation. Technical skills aside, domain knowledge and enthusiasm are both key ingredients of a successful programme.

Striving to retain knowledge and ownership of the data programme in-house are worthwhile strategic aims.

Either way, you’ll need to clearly define each role to both attract and retain the best people in a competitive market and ensure you hire the right people at the right time. Some of the worst job descriptions I have seen are for data scientists.

More planning equals less hiring of these expensive assets.

Whether you’re starting a new data programme or have struggled to embed an existing one, you need to make sure these 5 ToDos are in place. Otherwise you are going parachuting without the bag.