In Douglas Adams’ Hitchhiker’s Guide to the Galaxy series, researchers discovered that the answer to the Ultimate Question was 42. Only then did they realise that perhaps they also needed to know the question.
Questions are powerful things. An innocuous-sounding question at the right moment can expose gaps and flaws in an otherwise effective presentation. Just watch entrepreneurs crumble under the questioning in Dragons’ Den. Five minutes ago, they looked confident and in control. Now they’re practically weeping. But how do you find that ‘killer question’?
And this is, for many people, the key issue with analytics. If you don’t know what you don’t know, then how do you know what questions to ask to get the answers that you need?
Asking the right questions
After ten years of careful research and development, BeyondCore can now solve that problem for you. It claims to ask millions of questions of your data, and then analyse and present the patterns that it finds, giving you the story behind the data. It uses patented ‘smart pattern discovery’ software. Gartner predicts that this type of smart pattern analytics will be the most in-demand business intelligence platform in the future, not least because it makes business analytics and intelligence available to the ordinary business user.
BeyondCore has four main options for its charts. These are descriptions (what happened), diagnostics (why that happened), predictive models (what will probably happen next) and prescriptions (ways to make it better). BeyondCore’s CEO, Arijit Sengupta, suggests that BeyondCore V Beta has a unique set of capabilities, that make a data-driven world seem a reality, rather than a fantasy. These are:
- Being able to ask millions of questions following one single click;
- Providing explanations for the patterns in simple language that anyone can understand, meaning that companies no longer have to hire experts to understand and interpret their data;
- Using statistical techniques, it can highlight the patterns that are statistically significant, and remove those that occur by chance, or where the correlation is low-quality, from the analysis. This means that you won’t be left wondering whether a trend is real or not, or misled by low-quality data;
- Describing the reasoning behind predictions and diagnostics, so that you understand it, which makes the insights much more likely to be acted upon. Very few people will act on models that they don’t understand. The software can even explain the reasoning out loud, in a text-to-speech option;
- Enabling you to overlap your own unique insights into your business and influence the analysis and the automatically-generated commentary on the graphs and charts. This means that it can take into account one-off events that don’t show up in historical data, and you can also make it more relevant to your team by including references to conversations and discussions;
- Reducing the time needed to obtain insights by speeding up the processing time significantly.
What’s more, you don’t need any specialised hardware. You can deploy it on premises or in the cloud, and there’s no installation or customisation required.
Why does this matter?
Having an app to help you do something does not mean that you can necessarily do it. Say you have Powerpoint. That doesn’t mean that you can create or give a great presentation. It’s the same with business analytics. Giving business users access to data, and the tools to analyse it is a step in the right direction. But it doesn’t make them all analysts. Sometimes all it does is make them confused. There are just too many questions.
But BeyondCore’s pattern detection software gets past that. Instead of you having to ask it the questions, it asks millions of questions automatically of the data, and works out which ones are important. Effectively, it is an automated analyst. It will crunch all the numbers for you, work out what you need to know, and then walk you through the slideshow to explain the analysis and highlight the key insights.
BeyondCore won’t work for every company, not least because it needs to start with fairly good data. The data have to be semi-structured, and there must be at least 10,000 rows of data, which makes it better suited to larger companies, or those with a high volume of transactions. But in a world where data scientists are expensive, and everyone is trying to reduce costs, this ‘automatic analyst’ could be worth its weight in gold to provide insights to business users without them having to turn data scientist overnight.