Executives are failing to get any return on their Business Intelligence and analytics investments says Dr Rado Kotorov, CIO, Information Builders.
Speaking to CBR, Kotorov said that he had been asking executives during Gartner summits, with the CIO saying that the consistent answer was that no return is being seen.
Kotorov, said: "We have a lot of investments made to that area, every year we replace tools from one end to another and we are not seeing the returns."
What people are starting to realise is that having analytics is not enough, it is necessary to put them into the hands of the front line workers that make the decisions.
This isn’t a particularly new idea; Kotorov cites Herbert Simon an American political scientist and economist who won a Nobel Prize in decision studies. Simon identified in 1978 that the focus in the enterprise should be on the worker.
This is because the entire performance of the organisation depends on the individual worker, a concept which the IT industry has not taken into account.
Kotorov said: "We as an industry have not taken into account how to give information in the right format to the worker at the point of decision making.
"What I’m seeing is a trend that the operation people; the VP of operations, the president of operations those who have a really strong responsibility, they are looking to BI and saying I have thousands of dashboards and they haven’t helped me a bit to make operations more streamlined."
The change that is coming is one that will see information from analytics impact the bottom line.
Factors that have created this problem include technology not being scaled down enough for simple use on a tablet, because it hasn’t been thought of and because businesses haven’t necessarily wanted to enable their staff in this way.
Kotorov says that traditionally the BI and analytics space has been completely disconnected in the decision side, which is a problem that has often been raised by analytics companies when pitching their products.
Democratising data and extending it from being just the playground of data scientists and putting it into the hands of the everyday user is the goal.
The question is then raised, how many companies are being successful in putting analytics into the hands of their workers?
Clive Longbottom, analyst and founder of Quocirca told CBR that this is not being achieved with much of the problem being the expense of the licenses: "If they continue charging £300 plus per user or in some cases a lot more – then the cost is completely prohibitive."
To combat this there is a necessity to have different license types, one so the knowledge workers or workers that only need to see the data can have a free license and enterprise licenses for those that really want to play with the data to discover insights.
Longbottom went on to say that what you don’t want to have is task workers with the power to make data decisions when they don’t understand the data.
Although analytics in the hands of the store worker may be ideal for companies looking to sell products, it could be one step too far becoming tech for tech’s sake.
Kotorov uses the example of an application Information Builders has with Ford, which looked into the problem of warranty costs being higher than other automotive companies. It was discovered that some dealerships were replacing parts that could be repaired; the cost of repairing is a fraction the cost of replacing.
While this is a great analytical finding the question arises whether or not you can make the mechanic do this analysis on a per part basis.
Kotorov says the answer is: "You have to do this in an operational system so they are diagnosing the car and see what the status is and the system provides the feedback you need."
However, this means that the mechanic needs to spend more time doing this which is opposed to the incentive of mechanics to move the job along quickly.
A mechanic doesn’t have time to sit and deliberate the decision, he says and the same is happening in the fast food industry.
Although by bringing analytics to operations you are opening data to be used more quickly, there is still a time delay compared to the 30 second decision that people have been making in their heads for years.
The answer to this is to take away the decision making process from the mechanic and staff in the retail store, which in theory means that data can never be truly democratised.
The problem is that people will have to take extra steps in order to come to an outcome, while this may work fine in non time sensitive industries, when time is of the essence will people take time out to look at the analytics or will they trust their gut?
Longbottom disagrees that there should be a cut off point to who has access to analytics instead he thinks the path to democratised data analytics could be through guided analytics.
Longbottom told CBR: "I disagree that there should be a cut off point. You can have guided analytics. If you look at franchise garages like Ford, they would plug into an analytics device on the car (OBD reader) doing that can give you enough information and get in a position to guide the decision, like the Watson approach.
"You can see there is a part in a bad way but it could be fixed, can then look at data and make a decision, the task worker level in more of a position of guided analytics."
One company that is looking at democratising big data through the widespread deployment of data analytics tools is Logi Analytics, the Reading based company is aiming to enable even the average user to run analytics.
As previously mentioned, there is a price problem associated with the ability to give analytics to every user.
If you have a company of 40,000 people then £300 a license makes for a daunting price tag that most companies would be put off by.
While the idea is great, the reality of it is that it may not be currently possible. There are options out there such as Tableau which has free software, but they may not fulfil the needs of every user.
Questions also need to be asked as to why your company should be doing this, rather than simply having the tech for the sake of it.
It’s important to not go about it the wrong way and to ask what they are trying to answer, who really needs it and why.
Giving the average worker a SAS dashboard may give them every tool they could need but they are unlikely to know what it means.
While deploying SAS for the data scientist and Qlik for the knowledge worker leads to its own complexity.
The complexity posed by this is that of keeping the two in sync, so the importance is finding a single engine that suits all needs to all levels, which is easier said than done and has the potential of tying you in to one vendor.