Danish physicist and Nobel Prize winner Nils Bohr once stated, “Prediction is very difficult, especially if it is about the future.”
This does not stop companies from trying to use data to model what will happen in the future and where they can improve their chances of success in the market.
In its Worldwide CIO Agenda for 2015, IDC predicted that by 2018, 30 per cent of CIOs will have rolled out a pan-enterprise data and analytics strategy.
From a data perspective, this kind of investment involves more ways of handling information and using it to look forward into what may happen to the business and the market that it works in.
From a strategic level, using analytics – and most likely predictive analytics, to be specific – and techniques within the business means that everyone across the company has to have access to data and be more skilled in using it.
Prediction #1 – How management teams use data will change dramatically
For the management team within a business, this represents a big change in how they look at their company and gauge whether it is being successful.
Rather than looking at data on a monthly basis, reports on company performance can be put together weekly or even daily. What is still unclear is how this will impact the decision-making process at board level.
For CEOs, the ability to get data every day can help measure performance and lead to identifying where opportunities lie, but does it fit well with the long-term view and direction setting that they are supposed to be responsible for?
On the board side, there is still a gap between the board-book style approach for gathering data and displaying it against the real-time data analytics that the company uses in its day-to-day operations. Should non-executive directors get access to this real-time information too? Will they be able to make use of it in the right ways?
Providing this kind of management dashboard involves understanding how to balance the needs of those two groups.
For the CEO and other members of the leadership team, getting down into the details with new metrics like lead-to-cash and lifetime customer value will be the most useful way for them to make use of data.
For board members, looking at other metrics can be just as useful to see that the company is being run well.
Prediction #2 – How companies put together their numbers will evolve
The biggest challenge here involves getting all the necessary data from across the business linked up from multiple sources, and then getting this into a form that is usable for decision-making.
This process involves more than just letting people analyse information with data discovery tools. Data discovery is great for that initial analysis in order to test a theory, but each individual will have his or her own data sources and ways of combining information.
If all these data sources are not centralised and managed, it leads to discussions about whose data is “right”, rather than what the data can actually show. For businesses that want to create competitive advantage through use of data, getting this consistency of approach is critical.
Looking into the future, this process will involve much more integration of information within the “data tier” – where the information lives within applications or as part of wider data warehouses - and getting this into a “user tier” that everyone can have access to.
Splitting this complex process into two specific sections should make it easier to manage all the different data sources centrally, while also giving more flexibility to the users that want to analyse and play with that data.
Linking up data centrally is nothing new – it’s why data warehouses were invented in the first place. However, these centralised data sources are not going to be enough for people who need information from lots of different sources. This is particularly true if you are looking at data being used in real time for decisions.
This links into how data is getting used for decisions faster than in the past. The issue here is that more data is getting collated and captured; at the same time, the fact that more information might be available means that there is more reason to store it.
The advent of big data solutions like Hadoop and Spark, along with the volume of data that NoSQL solutions can handle, does mean that these larger sources of data can be used, but the ability to link this all into a decision is still relatively new and untested.
Companies are currently working at a cadence where data is being given to them on a regular basis rather than a real time one.
Prediction #3 – Functional data analytics expands to the end-user
In 2015, we will see more breakdown and cross-over between use of data in functional areas alongside the markets that those companies are in.
More sales and marketing teams are bringing in analytics as part of their requirements. However, some sales divisions handle commodity products and their approach to analytics focuses on how to improve order taking.
For other companies that have more consultative selling strategies, analytics will cover performance of specific channels and online. For the team behind the data science and BI within a company, it’s important to understand how the business runs its operations and the kind of process that is involved, as this will affect your approach to analytics.
Sales in one company does not work like sales in another – even within the same industry, companies will sell in different ways, and the product mix may be different.
However, data can be used in more ways across the whole business. These have traditionally been broken down into roles like sales, marketing, finance and operations.
As companies look to make more use of data, modeling how that information is used and put together is something that tends to be done centrally. This misses out on all the potential that data has to make a difference to how a business runs itself. Improving performance should not be looked at in isolation.
The best way to think of this is around how much improvement you can make centrally compared to at the edges – would you prefer to improve one decision by 100 per cent, or 100 decisions by one per cent each?
The cumulative impact is much greater than the single improvement can deliver. Better use of data in marketing, for example, should also spread into other teams like sales and operations in order to help those teams improve too.
Getting data out to everyone is therefore a key differentiator for companies in the future. Bringing the right information together in a controlled way, but then distributing it across the business, has the potential to help everyone in a company make better decisions.
Prediction #4 – Predicting more predictive analytics
The past few years have seen discussion of predictive analytics grow again. Predictive tends to go in cycles, as companies seek to make improvements in how they manage their operations but have not yet seen as much benefit as they aimed to achieve.
However, I think predictive now beginning to fulfill its potential, based on two things.
One is the availability of more data to use as part of the prediction modeling in the first place – with more information on the success of campaigns like marketing projects and the ability to track online behavior, companies can seek to understand their customers more effectively.
The second is how the ability to create algorithms and data science has spread beyond the most in-depth enterprise environments and become supported more widely.
This “democratisation” of data and analytics – made possible by approaches like cloud computing – should help more companies be able to implement predictive analytics and see value from it.
However, it will remain the preserve of specialists unless companies start to implement predictive analytics in ways that serve a practical purpose across the organisation.
Making predictive analytics practical can involve some change to business processes and how people work. While not everyone in the business will have to be a data scientist in the future, it will be important to ensure that everyone understands the role that data plays in the first place.
Getting to this understanding around data is the responsibility of the CEO, aided by the CIO and / or the Chief Data Officer if one is in place at the company.
Now it’s important to put predictive analytics into context. Implementing this kind of approach is not magic that can charm money out of customers’ pockets or make processes more efficient by itself.
However, it can make a good company great and great companies excel at how they run their operations, helping them concentrate on what their key strengths are.
Southard Jones is vice president of product strategy at data analytics firm Birst (opens in new tab).