Demo: Go!

Become an early adopter of the 'As a service experience'


After a lot of coffee-fuelled discussions on how best to explain what is and how to use it when analysts need to understand what explains their customer behaviours, we decided to go one step beyond writing tons of posts and articles. And, boom! We've made a demo app of to do the job. We cunningly named it … Demo (I know, this was the output of a lot of brainstorm sessions!). Wait, you do not know the best: it is a free (Boom again!) single session, SaaS application available here:

We've gathered together some of core features for you to answer predictive questions more easily: detection and removal of useless features, sorting and display of meaningful features, combination of features to reuse a predictive scoring formula against new data. You know these things can save you hours of hard work when carefully automated. And, like any good thing in life, it has some added extras. We have inclued some data sets so you can get a feel of how to easily predict customer behaviour even without your own data set.

Why? Two reasons.

Firstly, because the proof of the pudding is in the eating!

We often hear from our users that launching a new predictive analytics initiative is like exploring new ways of working without really knowing whether it will lead to a positive impact for our day-to-day business. And by positive, we don’t just mean ROI stuff. We also mean being able to answer in two minutes that question from Jane in marketing about what drives conversion or hot to avoid crunching these dreaded pivot table on excel overnight. You get it, our users want to understand and predict their customer behaviours but are also wondering what are the variables that drive churn, are not clear how to reduce brand fatigue on their marketing campaign or are wondering how to assess the right customers to include in an upsell campaigns. Whilst these are key KPIs for them, they’re not able to take an informed decision before starting on whether it’s worth using predictive analytics or not. What they see, for sure, are hurdles like: my data is not clean enough, my team is not skilled enough, putting things to production is (suuuuch) long shot, reliability is key and hard to assess, I need a heavy infrastructure to store and compute, etc. You get it, these are some of the hurdles that remove with

Secondly, because we've listened to our early adopters: They want to get a feel for the product quickly.

And now can be tested as a SaaS version. With you can run predictive analytics in a few minutes and access them easily via the AWS marketplace. This removes the usual hurdles of provisioning, privacy, scalability, etc. However, our users are focused on the sales and marketing and some of them told us  they have been confused by the AWS element (even if the AWS marketplace is a great mechanism to deliver innovation). It makes life easy but, understandably, AWS is sometimes not part of their day to day ecosystem. So we decided to go one step further, delivering an "as a Service" demo experience.

At PredicSis, we are dedicated to removing the hurdles that data savvy companies and users face when they run analytics initiatives. Companies use because they want to avoid the often big gap between business users and analytic team . They want to increase interactions between these teams, reduce the time between a predictive question being posed and its answer (to minutes where possible). They tell us they were frustrated to talk about "big data" and rather want to runlots of small analytises continuously. We have even came across an article summarising how users may use, introducing the concept of Minimal Viable Prediction. Sometimes, when we talk it through, we are even being told it sounds a big magic! But that’s just what auto ML enables: a growing field of machine learning that we are aiming to democratise. So, to democratise it even further, what could be better than giving users a sense of the experience via a Software as a Service demo?

So, here it goes, it is all yours to discover now: discover meaningful and reliable insights you can understand and use, with a few clicks, with or without any advanced data science skills.

Once signed up, as the 2-minute video shows, in just few clicks you will able to:

- create projects
- upload some datasets of yours, or pick among in-app ones
- define which column is containing the variable you want to understand (yes, we’re talking about supervised learning)
- select / unselect descriptive features to include in the analysis
- have run the analysis
- discover insights, filtered and sorted, explaining the outcome
- assess the predictive power of those insights

And, if you have questions or want to share something with us, get in touch in-app or give your views here. We value your feedback, thanks in advance!