Adapting to the environment or to the user’s needs

Or how Darwin evolution theory applies to an analytic start-up

Predictive analytics and Tableau

I recently read Erik Vermeulen’s interesting post about the different ways for start-ups to grow in a digital age and he mentions three specific considerations: building an eco-system, leveraging technology to deliver constant innovation, and adapting to the environment. In his article on hackernoon (click here for full details) Erik uses the example of an Indonesian start-up, Go-Jerk, which he says is definitely NOT a ride sharing company, or even a company in the conventional sense, despite appearances (again, I strongly suggest you read the article as I do not want to do it an injustice here). At PredicSis, leveraging technology to deliver constant innovation is what we are all about - through helping our clients to understand and predict customer behaviour; but it was Erik’s third point in particular - adapting to the environment - that resonated with me for our own particular set of start-up challenges. We constantly evolve and made a substantial move early this year by enabling our users to provision the service without any hassle via the AWS marketplace. This is something that our CEO has already explained in a previous blog as it made him very happy :)  . But there is a new evolution that we have noticed in how our users use PredicSis which we have noticed and are starting to act upon.

Let me explain.

As mentioned briefly above, PredicSis is a start-up used by large and small organisations to profile and predict customer behaviour. Sales and marketing teams use to understand what drives churn, fraud, lead conversion and to help them be more effective in their marketing activities. So far, so good.  It is something known as Predictive Analytics, or predictive Business Intelligence, and is a growing field that complements the more established field of B.I.

So where’s the challenge, you might ask?

Well, the reality is that predictive BI is still a nascent field, whereas BI is very well established, with well known organisations selling a range of packaged solutions. These include the likes of Qlik, Tableau and others. But it can take significant time and effort for analysts to sort, group or discretise the data to find the relevant information, before being in a position to deliver a crystal clear analysis. This is a pain point for them. Take the scenario where a key stakeholder who is looking to understand some data, say why are some customers churning or which customers to include in an email campaign, pops their head round the analyst’s desk at 4pm and asks for some insights for the following day. Gone suddenly are any plans for an evening chilling out….

But it doesn’t need to be this way.  Analysts in organisations that use PredicSis have worked out they can use the PredicSis auto ML capability to automatically extract information from the data to accelerate these BI processes.This phase of information extraction is called Data Enrichment, as identifies meaningful aggregates or features. It does not entail adding additional sources of information but rather squeezing the data to its core, to extract optimal meaning. At its most simplistic, the workflow now resembles something like the one below:

BI Tool predictive analytics on AWS

A video is worth a thousand words...

We’ve taken a simple public data set and come up with an example and filmed a video. The analyst in this example wants to identify the profile of US citizens earning $50k or more. It shows that even a simple data set requires some time to be clearly understood. PredicSis cuts down the effort required from hours to just minutes.

What does this have to do with Erik’s third point?

The key thing for us is about adapting to the environment in which we are evolving. Pretty much the evolution theory of Darwin where one has to adapt to survive and thrive. The environment we are in is data driven and therefore very close to BI, one would think. BI is a well established field and millions of organisations around the world use it. In a go-to-market situation for a company like PredicSis, it is not surprising then that we have focused on predictive BI - our core offering. Yet the still nascent market of organisations actively using predictive BI or even being aware of it, is a very limited one. This has made us realise that we have an opportunity to adapt to the existing environment while at the same time drive adoption by adapting our message to the large base of BI users to take them on the journey from BI to predictive BI. These users, although now comfortable with BI, feel predictive BI is complicated, and are worried it will require a big investment of time or resources. Our job is then to facilitate the process whereby they understand they can use PredicSis to accelerate tasks they already do. And then bring them on a journey where they realise they can use the same data they already use for BI, drag and drop it into a csv file and then witness how PredicSis can move them from seeing only past behaviour to actually predicting future behaviour.

Intrigued? If you want to get a sense of, we made it easy for you through our demo page