Analysis is deeply embedded in organisations; all accountants expect some level of analysis from you, even if it is as simple as “Total Monthly Sales”. The open secret of the analytics industry is that those “What happened” or Descriptive Analytics will never go away. In fact, they will continue to make up the vast majority of what you and your organisation report upon. Predictive analysis and prescriptive analysis seek to minimise the difference between forecast and reality, while prescriptive analysis also aims to give some scientific basis to forecasts.
The main benefits of modern analytics platforms can be realised using an entirely “Descriptive” mindset… The Analysis Delay (as I have taken to calling it), is the maximum time delay between an action and it is appearing in reports. Ten years ago, when I worked for a FTSE 100 company, our monthly Analysis Delay was up to 35 days. The transition to an early version of Power BI brought that down to 1 day. If that alone does not come across as powerful, let’s consider me. This automated change freed up nearly two weeks of activities for me per month directly, and a further week lost to AoB items from the board – “Wait”, I hear you say “, AoB items, how were they reduced?” This is a more subtle one, but the second thing Power BI brought in was it made us standardise. “Official” data was available after the overnight refresh so why would you not use it? Everyone used the same Semantic Model, and that being refreshed and updated daily meant that it was possible to “investigate” what would traditionally have become an AoB from board meetings there and then (we even stopped extracting screenshots for the PowerPoint deck and instead went through a specific report).
Bob Cratchit is everyone’s hero; in our story, he is a much-maligned analyst, constantly working to produce the reports while also manually cleaning and preparing the data. It was no wonder his laptop – Tiny Tim – was struggling to cope; offsetting and automating the cleaning and preparation work into the Power BI Service will mean all that Bob is doing is configuring the Semantic Model and then the visualisations. A standardised Semantic Model will allow Bob to implement secondary reports to manage, evolve and grow a better solution. Continual Improvement is on the table.
Carol’s Cabs doubled down on automation; no doubt, once approved, they would also switch to self-driving cabs. Predictive Analytics would talk about “Maintaining performance in Manhattan” or “increasing cab revenue in Queen’s by 5%”. There is little evidence as to why that is the forecast. The Prescriptive Analysis that we suggested for Carol’s Cabs would focus more on other external factors, looking at weather and events as a simple starting point. It would be possible to extrapolate an expected Cab requirement for any given time by looking at what events are going on in a particular area and the forecast for the weather (another topic where predictions are in high demand). These two variables are fed into a model, along with many others; that model will, for example, suggest that Carol can expect 27.5 journeys from Grand Central Station in the hour before a Basketball game at Maddison Square Garden. That number will increase to 39.7 if the air temperature is below 5 degrees Celsius. However, if the air temperature drops below 0 at the Crysler Building, the expected utilisation drops to 17. There is too much going on and only pseudo-science here. Still, your prescriptive models allow vast complex interactions to be identified where the correlation can best be considered an un-chartable 20+ axis linear regression. Remember that just because you identify correlations (Temperature at the Crysler building) as being there, they are not causal. In our example, I would imagine this indicates a point either where the main “Manhattan Temperature” is taken or that people do not want to stand on suburban train platforms in the cold or even that the suburban rail service is unreliable when things start to freeze, the only thing Carol would need to know and monitor would be if that correlation holds, does she need to know the specifics, especially considering the reality is that all three are true.
Now running my own business, Geordie Consulting, I want you and your company to excel and to get the most out of your data. We believe that only if your company succeeds can we succeed. So, in 2024, why not invest in yourself and your business by arranging a call with us? We can quickly help move you forward with our support or without it. There was nothing inherently wrong with Ebenezer’s business; it was more than a couple of little tweaks, which is more and more the situation we hear from our new clients. We can help steer your internal team in the right direction with a light touch or offer a more Centre of Excellence as a service approach if that is the most advantageous for your business.
So have a Merry Christmas and a Happy New Year from all of us at Geordie Consulting
