Since October 2017 the pace of development has really ramped up already 2018 looks like it is going to be another bumper year. The focus of the year is very definitely going to be about the transition from old fashioned Adaptive Intelligence – I prefer “Adaptive Intelligence” to Artificial Intelligence.
Two main topics currently seem to be dominating my world.
- When will BI be declared dead!
- Which platform should I be betting the farm on?
Death of BI
The statement that BI is dead or dying is in itself an emotive once and one that is causing a great deal of conversations within the field at present. The key is to realise that this is not what is actually meant by what is being said. To understand you have to appreciate the three flavours of analysis, metrics and visualisations; Descriptive, Predictive and Prescriptive. Ultimately in a world of Big Data we need to be transitioning away from Descriptive Analysis (what happened) and moving towards predictive (extrapolate performance targeting and base-lining) with an ultimate goal of prescriptive metrics (extrapolation recommends action/activity)
Now if you understand that there are different types of analysis you can see what is meant by the death of BI. Descriptive Analytics are dying and the expectation of the audience is not just that a BI department/team/person can tell them what has happened, but that they can be put in a position to estimate what will happen with a goal of being able to be certain of the future to a defined percentage i.e. 80% chance that revenue will increase month on month if product is launched at a specific price point. The demand of now is for a team to not become a bottleneck or to deliver performance figures weeks or even months after the event – analysis must be available as close to realtime as possible and issues must be able to be identified quickly and efficiently ideally in time to influence performance.
overall I am not worried about the statement that BI is dead more I see it as a challenge to myself and my team to make sure that our involvement is more upfront, so building data analysis tools that automatically refresh and update to the “delay” or Analysis – Fact gap is minimised. It is a luxury to be able to have a week or more now between performance and a pack being available, commercially data over a day old can be considered “stale” so the expectation is to have more sooner now. Across 2018 I expect this to continue with more and more machine learning projects starting as we try to work out how we can move from a predictive model to a prescriptive. I honestly can’t wait!
Between 2016 and 2017 the Business Intelligence world took a huge step forward
Microsoft almost re-entered the BI arena after some time in the wilderness re-developing their solution from the ground up and Tableau continued to grow. The result in 2017 seemed to really be a battleground forming between the two platforms. I have lost count of the number of discussions or even arguments that I have had in the last 12 months over which platform is best and which one will deliver the most for a business. The truth is any platform in the quadrant will deliver a good if not great result, Business Intelligence (all three types) are not made by a tool, they are made by understanding what is going on and how to report it. It is entirely possible for a company to deliver interactive BI with zero software costs beyond the core OS and Office (and not even Microsoft). A tool will only help in terms of simplification of the process and improving the aesthetics of your visualisations. In 2017 Tableau and Power BI broke away from the pack and became real stand out leaders in the field leading to a whole host of comparisons and arguements for or against either platform. For the sake of argument I am only going to look at these two, but I would love to hear your feedback.
Tableau and Power BI both aim to deliver dashboards and visualisations and – in my personal opinion – both do so extremely well, there are however differences. Data Import is something that Power BI does better while story telling with data Tableau does better. Again these are my opinions no one elses. From my experience working with both I would think of the following factors.
- Who needs to look at the report?
- Do the consumers of the report need a single static view or a do they need to interact with the data more?
- Is this a report or an operational aid?
I’ve found a few more questions, that will further help, but I find those three questions really help to narrow down the tool that will work for your use case. The truth however i that for larger clients both will be right with a seniority differentiator between the platforms i.e. most staff access reports via Power BI with a focus on operational dashboards and reports that can drive actions, these reports are then mirrored into Tableau (by importing Power BI data models) to then have the top-level board reports produced in Tableau.
I’ve given some simplistic highlights for selecting a platform, but simply put if you have limited budget start with function, reporting must drive operational activities i.e. Sales in area x are down so we can do something about it sooner rather than later. Showing that only to a top-level so executives know your sales are down in a specific area before your staff do only serves to alienate your audience, so target operational reporting as the priority over the executives – I know this is easier said than done especially as those executives are often who you need to sign the cheque.
Good luck and I hope I’ve helped and not muddied the waters too much.