Data Challenge

Australia’s Data Analytics Challenge

As companies enter 2021 with a focus on returning to some form of normality, the strategic momentum continuing from 2020 into this year is the importance of data and digital. 

However, in spite of many companies upgrading their data and digital capabilities, much opportunity has not been realised and organisations we are talking to, especially the people in the senior data and analytics roles, are deeply dissatisfied with the current state of affairs.

They talk about data and analytics being paid much lip service but little or nothing happening on that front or at best once off activities to “tick a box”. So the question is, what is holding Australian companies back?

To answer this question, Biarri and McGrathNicol went out and interviewed senior executives in data and analytics roles across a variety of organisations in Australia. Our findings and results are presented in this video.

If you’d like a copy of the deck or any further information, please reach out to Evan Shellshear as shown in the video or use the contact form below to get in touch.

Get in touch

  • Hidden
  • Hidden
  • Hidden
  • This field is for validation purposes and should be left unchanged.

Biarri Workforce New Feature: Updated Shift Swaps

At Biarri we understand the challenges that Workforce Planners and employees face when rosters are published. We understand that circumstances outside of work may change and affect the ability for an employee to complete a shift, conversely, Workforce Planners’ and Managers’ pain point of finding a replacement to cover a shift or reconfiguring the roster to suit. 

We have listened, and Biarri is excited to introduce the new enhanced Shift Swap feature on our Biarri Workforce Mobile App.

Created out of necessity, employees can now more simply either Swap Shifts or Offer Shifts to other available employees.

When swapping a shift, we’ve removed the stress of trying to identify which employees are available. Instructed by the rules engine, the Biarri Workforce App generates a list of compatible  employees to cover a shift. These compatible employees comply with rules defined in the Admin Tab, ensuring employees are compliant with skills & qualifications, fatigue and availability rules. 

See a step-by-step walkthrough of how simple swapping shifts has become.


Read more
Worker Fatigue

Modelling Patient Flow to Improve Service Delivery

Running a hospital is a highly complex activity. There is not one division with one need, however, hundreds of areas are highly interconnected and all rely on each other to deliver life saving outcomes. Patient flow in such a complex system is a serious challenge.

In 2019 alone, there were over 11 million separations in Australia (a separation from a healthcare facility occurs anytime a patient leaves because of death, discharge, sign-out against medical advice or transfer, it is the most commonly used measure of the utilization of hospital services). These were just the patients who were discharged, let alone those who may have unnecessarily visited a hospital and sent home immediately.

Managing these millions of patients and providing the high level of care we have come to expect is extremely difficult.

This is why a Victorian Hospital recently turned to Biarri to get some help with their patient flow challenges. Biarri turned its mathematical expertise on the problem and what came out of our work was a great result with substantial improvements and clear benefits to the most important person of all – you.

The Patient Flow Challenge

The hospital approached Biarri after realising that its emergency department was consistently overcrowded at seemingly predictable times and for seemingly predictable reasons. The staff and executive team at the hospital knew that they could better plan and manage their overcrowding problem, and we at Biarri knew that we could provide them with the tools with which to inform those decisions.

One of the key metrics that quantifies the overcrowding of an emergency department is the NEDOCS score. In our case it wasn’t that the NEDOCS score didn’t provide value, it was that the staff were not able to anticipate when the score would rise or fall. Both management and clinical staff felt it should be possible to better predict when these spikes would occur, and therefore better prepare for them.

So the business problem for Biarri became, can we produce a tool that is able to accurately forecast the bottlenecks and queue lengths within an emergency department?

It was this fascinating challenge we solved and have developed a unique tool specific to this hospital.

The Approach

Biarri’s approach to this challenge was to build a custom simulation model that would act as a digital twin for the hospital’s emergency department. This allows the staff to simulate how the patients move through the different streams of the emergency department. The simulation model is built upon three main objects: 

  1. agents, 
  2. processes and 
  3. resources. 

Let’s define these relative to the emergency department.

Patients are the agents that move through the system. They possess attributes such as scan type and scan priority that indicate how long it takes them to move through the system. Agents go through processes such as triage, scanning and reporting, which move them along the emergency department. Finally, doctors, nurses and the machines they use are the resources that are required to move our patients through the processes.

All the quantitative values, such as the frequency of arrival, or the time it takes for a CT scan, etc are found using the historical data from the emergency department. Using the past data is the best way to get a digital twin that emulates the workings of the emergency department accurately. This was combined with some smart twists on existing tools such as Discrete Event Simulation to simulate a high fidelity model able to accurately model the emergency department.

The Patient Flow Solution

At its core, Biarri’s patient flow tool provides users with the information to better understand how the future of their emergency department looks like. It does this by providing staff with feedback on the development of patient queues and bottlenecks at different stages of the emergency departments process.

Given the nature of simulation modelling, the tool provides a distribution of future queue times, with confidence intervals to encapsulate the likely ranges expected in the future. This provides insight into the changes over time of queue lengths at different stages of the emergency department. Not only does it give a general overview of the emergency department, but a more granular look at each queue.

A big plus for the patients is that from this queue length we were able to derive expected wait times so that patients are now able to know how long they expect to wait based on how busy the hospital is.

The tool also allows for scenario analysis. This gives hospital staff the agency to compare how their current conditions compare against changes in rostering when it comes to influencing patient flow.

With this solution, Biarri aims to give hospital staff the information to make sound decisions around the rostering of their resources and the structure of their emergency department.

What’s next?

Biarri’s new patient flow modelling tool was a great success and has proven its ability to accurately model patient flow and provide timely and accurate warnings of future bottlenecks.

The tool is tailored to the Australian environment and now having been demonstrated on a Victorian hospital, Biarri is looking to roll it out across Australia. If you have patient flow challenges and could benefit significantly by being able to better plan and allocate limited resources, then get in touch.

Get in touch

  • Hidden
  • Hidden
  • Hidden
  • This field is for validation purposes and should be left unchanged.

Child learning at desk

What is a dollar of Education worth?

Every year, over 16% of the Australian population enrol in schools across the country. Some of these students choose to pay for their education, whilst others don’t. Inevitably, a discussion will arise regarding whether there is any real academic advantage in a more expensive education; the age-old private vs public debate.

Due to the swift rise of data science techniques and resources over the previous decade, an understanding of this problem through data science is now more attainable than ever. This article seeks to assess the relationship between socio-educational factors and the outcome of a students’ academic achievement and, most importantly, to determine whether paying more for schooling is justified through higher academic outcomes, in particular those measured on standardised tests. We achieve this through a deep statistical analysis on both the 2016 Census, and 2018 NAPLAN datasets.

The results reveal some fascinating facts about what seems to matter for education and what doesn’t. 

The Education Problem

The goal of this research is to understand how socio-educational factors affect student academic achievement. Taking the NAPLAN data, and purely considering a school’s average score across each of the six tests, we find the results presented in the following figures.

Figure 1: Average NAPLAN Scores in Writing and Numeracy, across all schools in Australia (2018), categorised by school type. At face value, independent and Catholic schools appear to outperform government schools in both domains, but the understanding is not that simple. The bigger idea of why these schools perform better represents the aim of this article.

From the above figure, independent and Catholic schools do perform marginally better when simply compared to government schools. We could take this at face value, finish the problem here and say, yes, paying more for schooling and going to a private school holds more promise of achieving a better result. 

Unfortunately, drawing this conclusion is not realistic. Quite simply, comparing schools based totally on their results does not paint the full picture. A multitude of other factors may ultimately affect how a student performs at school, like a student’s upbringing, a population’s demographic, or a school’s characteristics. Accounting for these, we see a very different response to the problem, that is, school sector has a relatively insignificant impact on students’ performance.

For a concrete example on understanding how other factors may affect student performance, take the ACARA defined ICSEA value, best explained here. The ICSEA value tells us a variety of interesting things about a school’s student-base and their socio-educational position. If we illustrate the distribution of ICSEA scores across each school type, we find the results shown in the following figure.

Figure 2: Distribution of ICSEA across different school types. ICSEA represents an index of a school’s socio-educational position, indicating collective student advantage irrespective of their school. Clearly, there is a vast difference in socio-educational advantage between the sectors, which this article seeks to quantify.

In the above figure, we see a much wider spread on ICSEA scores for government schools, and higher averages achieved by Independent and Catholic schools. Based on this, it is a fair assumption to say we might expect Independent and Catholic schools to, on average, enrol students from more advantaged socio-educational backgrounds.

If students from a more advantaged background on average perform better, then by extension, we can assume that independent and Catholic school results are influenced by the students they enrol. So naturally, we would see independent and Catholic schools performing better than government, based on their student demographic. This simple intuition can be extended to many other factors, and represents the need to account for all possible variables when looking to truly understand how tuition fees affect student success.

Approaching the Problem

Our approach takes the NAPLAN results from 2018, combined with the ABS 2016 Census data, and uses these to understand how different socio-economic factors affect student performance. The main problem we’re trying to understand is enormously complex, with many variables that could ultimately affect academic achievement. In particular, we’re looking to understand the following:

  • A student’s background, i.e. family encouragement, parent contributions, extra-curricular activities, etc.,
  • Contributing factors to school fees, i.e. how tuition fee depends on school size, location, and reputation in a community.

Naturally, these include a huge number of possible variables, the challenge being that some of which are particularly difficult to quantify. For example, how do we express a student’s natural academic ability as a numeric value? This question is a foundation of the main challenge the problem presents.

Figure 3: List of possible factors that might affect student academic achievement, categorised between school demographics and school factors. Note this is a subset of an inexhaustible set of possible factors that may influence student achievement.

Considering this, our final solution needs to account for high levels of complexity, whilst making do with the limited data available. We took advantage of two core approaches:

  • Making assumptions to reduce the scope of the problem. Notably, to limit the effect of confounding variables, the scope of the project was narrowed down to exclusively include Melbourne schools, within grade three and five student results. This controls for background education effects which would be more prevalent if we analysed later years (i.e. after completing primary school).
  • As we don’t have the luxury of being able to carry out randomised controlled trials, we used proven and well understood statistical techniques to establish the associations between variables. We implemented our solution through Instrumental Variable analysis, a form of linear regression that can account for unobserved variations in data. This analysis ultimately gives us a means of accurately modelling the problem using a restricted dataset. Additionally to ensure the coming results are as robust as possible we have applied a number of techniques to account for confounders, collinearities and other problems which cause such analyses to go wrong.


The beauty of using a regression model is that it provides the ability to identify the strength and direction of correlations between each of our model’s variables, and school achievement. This is characteristic of using regression models, and incredibly useful in our use case.

In simple terms, the model returns a single numeric value for each variable in our model, termed model coefficients. Given a one-unit change in each variable, the predicted student academic achievement will then increase/decrease by the model coefficient. For example, in a model that estimates a variable representing the percentage of female teachers in a school to hold a value of positive four, we can assume that a 1% increase in the percentage of female teachers at the school will equate to an average increase of four marks in the NAPLAN.

Figure 4: Illustration of model results. The box plot values indicate how each observation affects student academic achievement, while bolded values represent model coefficients. Values in the box plot are simply dataset observations multiplied by the model coefficient. Note the difference in how variables positively or negatively affect student achievement.

Using this method, we can summarise the model results using the illustration shown in the above figure, which shows for each variable how changes in these values in the dataset affect average student academic achievement. Although we used the average NAPLAN score as education output to build our model (instead of each subject area, writing, numeracy, etc), a similar result also applies to separate analysis on the different NAPLAN subjects.

Furthermore, if we take the average effect of the values shown previously, we find the result shown in the below figure. You can see tuition fee highlighted in red; clearly still positive, but providing a relatively insignificant effect compared to its counterparts.

Figure 5: Comparison of the average effect from each variable on student academic achievement. Notice the relatively insignificant contribution made by the Tuition Fee variable, as opposed to volunteer ratios, student/teacher ratios, and married proportions of the population.

What is a Dollar of Education Worth?

Most importantly, the final model coefficient for tuition fee comes to 0.001044 with this being statistically significant. Intuitively then, we know that a one dollar increase in the amount paid for schooling corresponds to a 0.001044 increase in a school’s expected NAPLAN score. In simpler terms, a $1000 increase correlates to a 1.044 point score increase.

We can investigate how the NAPLAN results truly scale to get a better understanding of what this means. The below table illustrates how to achieve different NAPLAN bands based on scores in the grade three and five Numeracy exams from 2018. Note that this information is publicly available here.

BandGrade 3Grade 5
Band 192.4 – 256.1
Band 2270.8 – 318.2
Band 3328.3 – 365.6195.5 – 368.6
Band 4374.4 – 417.9379.2 – 423.8
Band 5426.8 – 475.0431.7 – 476.9
Band 6486.0 – 682.5484.3 – 522.2
Band 7530.2 – 573.7
Band 8583.5 – 777.8
Table 1: Comparison of NAPLAN Scale Scores required to achieve different bands in the 2018 Numeracy Exam for students in grades three and five. The national minimum standard for grade three and five students is a band two and band four respectively. Note most of the bands range across approximately 55 points.

The national minimum standards for this exam are bands two and four for students in grades three and five respectively. To jump from the median band two score to the median band three score for a grade three student is approximately 55 points. A similar result holds for grade five students, and subsequent jumps between bands.

Therefore, we can assume that for a student to achieve a single band increase based entirely off paying more for tuition, they require approximately 55 additional NAPLAN Scale Score points. Based on our model results based on the discovered correlations, this is equivalent to paying an additional $52,000 purely in tuition fees.

In practice, the most students from Victoria will ever pay for schooling is around $30,000, rendering scores up to 55 points impossible. Median private schools pay around $9000 on tuition, and by extension, this would give approximately 10 additional predicted points. Obviously, these bonuses cannot solely drive student achievement into higher bands.

Figure 6: Comparison of tuition fee paid and predicted NAPLAN score increase. Assuming a 55 point jump is on average necessary to move to a new band, approximately $52, 000 is then required to do this solely by paying more for schooling.

By comparison, the median percentage of people in a population whose marital status classifies as married was 40.44%. With a coefficient of 1.104, this percentage is correlated with a score increase of 44.66. The maximum percentage of married people in a population used in this study was 51.61%, which would correspond to a score increase of 56.99. The interpretation here is that in tact families assist students with their academic achievements, which helps put the tuition fee benefits in perspective.

Similarly, the median proportion of a population who regularly undertake volunteering duties was around 15.25%. With a parameter estimate of 1.72, this means the median score increase related to the number of volunteer workers was 26.28 NALPAN score points. The maximum effect based on data used in this study was a population with 24.19% of people who volunteered, offering an increase of 41.69 points.

In comparison to tuition fees, it’s clear other variables provide significantly more influence in our final model. Ultimately, this demonstrates how in practice, school results are clearly influenced by many different factors, not just the school type or amount paid. This allows parents to make better decisions as to where to invest their money and whether living in the catchment of a school with high volunteer rates, married families who have highly educated parents could be a much better investment in their children’s education than paying significant tuition fees.


Our analysis and the results of it indicate that paying more for schooling does positively relate to academic achievement. Conversely, this effect does not provide substantial enough benefit to solely allow students to achieve outstanding results as was to be expected. Instead, other more important factors seem to encourage academic achievement, including family composition, a population’s level of education, community involvement, and a school’s structure. 

The key takeaway from this report reduces to the fact that these influences, namely family contributions and population characteristics, play a far more important role in determining student outcomes than the amount paid for tuition does. Notably, this aligns with the findings of Fryer & Levitt (2004) who undertook a similar regression style approach to understanding the racial test-score gap in early-year students. Notably, they show the following are important for academic outcomes:

  • A child has highly educated parents
  • A child’s parents have high socio-economic status
  • A child’s parents speak English at home

These results go further to enforce our findings that tuition fees specifically aren’t as important in contributing to student success. For further details on the analysis that lead to the results presented here, the reader is encouraged to download the reports and discover the limitations and depth of analysis presented there. The reports can be downloaded here and here.

What Next?

Understanding this, we need to acknowledge the complexity of this problem, and the limitations of what data made available to this project. Whilst these results go some way to understanding how paying for schooling may affect student achievement, a further study involving a more comprehensive dataset and deeper statistical analysis would be necessary to truly understand the full relationship between tuition fee and academic success. In particular, the results presented here are based on regression techniques which can establish associations but cannot guarantee causation.

We encourage future researchers to build upon the results here to establish the causes of the effects we see here.


Most importantly, the work presented here only investigates academic achievement at the school level. When deciding where to send a child to school, whilst still accounting for our results, utmost importance rests on considering a student as an individual such as:

  • What do they like to study?
  • What opportunities are available at different schools?
  • What will give the student the best compromise between doing well academically and making the most of what a school can offer socially?

Whilst mathematics might suggest one thing to be true, in practice it might not always be the case, a cautionary tale that can be extended to the world of data science and machine learning in the modern-day.


Fryer R, Levitt S. Understanding The Black-White Test Score Gap in the First Two Years of School. The Review of Economics and Statistics. 2004.

For inquiries, reach out to the authors, Nicholas Thompson, Meichen Qian, Jonathon Allport and Evan Shellshear.

Sales forecasting tool

Sales Forecasting Tool to Plan Accurately With Some Simple AI

Throughout my career I’ve worked with salespeople, as a salesman, and in roles supporting sales activities. Sales is one of the most important functions of a business as without sales, you have no business, no matter how great your product or service is. 

Sales is the fuel for any business to survive and thrive and this makes planning and forecasting sales one of the most important activities a business does.

That’s why running a business without effective and accurate sales forecasting is a bit like flying a plane without a fuel gauge. Of course, an accurate fuel gauge is not necessary or sufficient for generating or maintaining lift – the Wright brothers got away without one. But there’s a reason why modern planes have them – it gives pilots access to data to make the flight decisions to get from A to B. 

So how are you flying your venture?

Probably the same as most other businesses. 

You gather your sales team and ask them, ”how many sales will we have this year?” 

In the best case scenario, they review last year’s sales and make a guess based on gut feeling and intuition (which is not always wrong). 

Commonly enough though, a misalignment of incentives and company sales culture can manifest as a mismatch between targets (optimised for remuneration incentives) and forecasts (optimised for accuracy).

We can do better. 

And to do this we need to use data. But why is data so important?

According to the Professor of Digital Practice at QUT, Mal Thatcher, the 21st century will be the century where,

“By the middle of the century the only tangible asset on an organisation’s balance sheet will be data”

and this is true for your sales too.

To give you and your business a competitive advantage, we at Biarri have developed a simple, easy-to-use Excel sales forecasting tool for you. So it is time to become data driven now and with Biarri’s new tool this is extremely easy. 

Biarri has taken some basic AI techniques and put them into a spreadsheet that requires no macros, no plugins and nothing to install. The AI techniques in this Excel tool will help guide your sales team to make more accurate predictions for the coming year. 

You don’t need to be an expert in AI to leverage the tool. It does all of the hard work for you and provides you with data driven monthly predictions for the coming year based on quarterly sales patterns. You don’t need to know cutting edge AI to use the tool, just how to copy and base a small amount of data.

You can download the tool below for free. There is no need to leave your email address or anything. Biarri’s mission is to make the world more efficient via better decisions powered with mathematics and we believe this tool has the potential to make a difference for your organisation.

Your New Sales Forecasting Tool

Before you download the tool, it is worthwhile telling you what it is, and how to use it.

It uses historic data to establish a pattern and then extrapolates this pattern to be able to predict the coming year’s sales. 

Not only does the tool provide monthly predictions, it also takes into account quarterly sales cycles. Forecasting quarter-by-quarter aligns it with typical quarterly reporting and also captures the variance in quarterly sales. This quarter-by-quarter approach is designed for industries like retail which have some quarters with greater sales (e.g. Xmas). 

There is also a “bad month flag”. This allows users to indicate if something bad has happened in the past during months (e.g. COVID) and if similar events are predicted to occur in the future (in the PREDICTIONS tab). 

This spreadsheet comes prefilled with data to show you what it should look like. To use it for yourself, remove the data from the Monthly Sales column in the Data tab and replace it with your own data. The calculations and updates will be carried out automatically. All other cells are locked for your safety. 

How do I use the sales forecasting tool?

The steps to using the sales forecasting tool are as follows:

1. Collect exactly 36 months of contiguous sales data leading up to the month you would like to predict from. E.g. if you want to predict the yearly sales from January 2022 until December 2022, then collect the 36 months of sales data from January 2019 until December 2021. The model is set up for exactly 36 months of data, not more or less.

2. Copy this sales data into the Monthly Sales column in the Data tab (in green). The top most entry should be the oldest (e.g. in the example in 1., January 2019) and the bottom most entry should be the newest (e.g. December 2021 in the example in 1.).

Sales forecasting tool monthly sales column

3. In the Data tab now enter the first month for the monthly sales data in the month tab by choosing from the drop down menu (this cell is green). Also choose the year in the year column from the drop down menu (this cell is also in green).

Sales forecasting tool month selection

4. For each month, choose whether a bad event occurred (or not) by selecting Yes or No in the Bad Event column. If normal trading and fluctuations were occuring, then put No. Otherwise, if something truly unusual (e.g. COVID) occurred that significantly impacted your sales volumes, select Yes on the months which were affected by this (this column is coloured green).

Sales forecasting tool bad event input column

5. Once this data has been entered, go to the Model Analysis tab to see the outputs of the model. In the Predictions tab, if you predict that there will be any bad months in the future, select Yes in the corresponding months in the Bad Event column (which is in green). For this to have effect, similar Bad Events need to have occurred in the past otherwise this will have no effect.

Bad event output column

6. Finally, your predictions are shown in a graph in the Dashboard tab, with a table showing the cumulative results for each quarter.

Monthly sales prediction graph

Download the sales forecasting tool by clicking on the button below.

What does Biarri do?

Most companies begin with Excel sheets like the one provided here to start making once off decisions on key parts of their business. It is like the first flight of a plane with an often inaccurate fuel gauge caused by data issues. At some point, organisations need to lift up from Excel to correct, secure, easy to use and more powerful tools and this is where Biarri helps.

Biarri’s main value proposition is to help clients realise operational excellence in the way they run their business via AI. The core of this is excellent, data driven decision making. Biarri catalyses AI driven business decisions by using its cloud based set of mathematical tools, the Workbench.

To discuss how you can leverage your data and turn it into value to reach new operational heights, reach out with the form below now.

Get in touch

  • Hidden
  • Hidden
  • Hidden
  • This field is for validation purposes and should be left unchanged.

Markdown optimisation example

The Optimal, Best Markdown Rules During COVID-19 and Beyond

When my friends talk to me about managing a retail store they all enjoy it apart from one thing. They tell me that seeing the latest trends, working with cool brands and meeting some of the stars promoting the articles is great. It’s just that there is a recurring event that makes long days longer and they are sure can be done better – markdowns.

My friends imagine that a set of retail analysts spend days figuring out the optimal markdown to make sure customers buy the remaining goods at the highest possible price to maximise the possible gross profit. 

However, in many cases, there is no group of analysts.

Instead there is a fixed set of well-worn rules, year after year: first a 10% reduction, then a 30% reduction, and finally a 60% reduction until the product clears.

It’s artificial ignorance not artificial intelligence. The retailers have the data, it’s whether they use it – and can make use of it – or not.

As if typical markdown cycles weren’t difficult enough, COVID-19 has turned these processes into a nightmare. Companies are now stuck in a vice trying to clear stock but doing so in a way that loses minimal money and keeps the company afloat. 

The old rules are no longer working so companies are experimenting with new permanent discounts – 50%, 70% and 90%! Just whatever it takes to get the stock out the door before it’s too late…

What is needed, now more than ever, is a data driven, machine learning approach to suit each individual business. The best markdown rules use analytics.

Biarri’s research has shown that too many goods are too heavily marked down, even in a time of crisis like this, meaning lost profit margins. In addition, a one-size-fits-all approach doesn’t work – each business needs to take its own unique conditions into account for such markdowns to be successful.

The early conclusion? The best markdown rules

In the current situation, you need an advanced analytics tool (AI, Machine Learning, Optimisation) that takes into account the unique circumstances of your business and can answer questions such as:

  1. Each store has different stock coverage for an article, so how do we balance a single store’s need with a regional markdown strategy that mandates one markdown for all stores in a region? Or even more difficult, a national markdown strategy?
  2. If we allow markdowns to vary between stores, then how do we create a markdown strategy that prevents customers shopping between your different stores searching for the biggest discount?
  3. How do we accurately model price elasticity for goods with only a few sales each month and even fewer historic price points?
  4. When marking down, how do we pick a smart price point – an “anchor” – that increases demand much more than the surrounding price possibilities? 
  5. How do we accurately predict the demand for each individual article to know what its markdown sticker should show to maximise profit?

As well as much more.

A tool that maximises your gross profit needs to get all the above right for you and your business, not for a generic retailer in a generic retail environment – these disappeared a long time ago.

The way you resolve these challenges is by having an analytics tool which can:

  1. Group stores together using geographic and sales profiles to provide simplified markdown strategies which staff are prepared to implement.
  2. Provide a set of optimal markdowns, determined from the historic sales data, which can be applied equally across all stores in a similar area.
  3. Compute accurate price elasticities. Price elasticity is often like sausages; everyone likes them, but it’s better not to see them being made. Biarri avoids elasticity problems by bootstrapping data to measure how volatile the elasticity estimates are and then weights them accordingly.
  4. Cleverly choose price points based on known rules of customer pricing behaviour instead of just basing it on the markdown percentage times the price.
  5. Predict accurately the demand for an article at a given price point by not predicting a number at all – we predict a range.

The last point is key as merely having a “point” forecast is of little value, you need ranges of values so that you can meet your business needs to ensure that you will either:

  • Sell all stock with a high certainty at a good price, or
  • Ensure the shelves are never empty and you don’t leave customers with a bad impression.
Best Markdown rules forecasts

How do I use it?

So now that we’ve identified the key criteria to create the optimal set of markdowns to maximise profit but also clear out the goods, the final step of a successful markdown strategy tailored to your business is building trust in it. This means rolling out the new AI created marks in a smart way. If you roll it out across your entire range and there are any issues, the entire project will be doomed to fail – even if it was immensely successful on some articles.

The way to make this succeed is to do some simple tests in a scientifically correct way. Create a control group where you do what you normally do and test it against a treatment group with the new optimal markdowns. This can be done either on a store by store basis (A/B tests), but if that is not possible then it should be done across similar product categories so we can draw valid conclusions from the experiment to find the best markdown rules.

What are the benefits?

If instead of using intuition and “What we always did” to markdown products, you turn to analytics to do this smarter, the gross profit improvements can be significant. This technique has been the success story of a number of giants such as Walmart and Target Corporation in their retail divisions. Not only has Biarri seen significant benefits from using optimal markdowns, studies have also concluded that it is possible to improve profits by up to 20% by using analytics instead of gut feelings.

COVID-19 is affecting everyone, irrespective of industry, however the retail industry is one of the worst hit. People are scared and not visiting stores leading to a drop in foot traffic of up to 50% and supply chains malfunctioning with over half of retailers impacted by supply chain issues having too much of some stock (sporting jerseys) and not enough of another (running gear). 

So something smart needs to be done – and right now. Analytics based markdowns are exactly what are needed to turn the retail industry around. Off the shelf tools don’t cut it for a situation like this as they are not customised to cope with so-called “out of sample” events like the current coronavirus crisis.

So reach out now to find out how you can clear out unwanted stocks effectively and maximise profit at the same time. Biarri’s tools are simple to use, customised to each customer and driven by accurate and correct mathematics to ensure the best result. They provide you with the best markdown rules every time.

Better yet, to simplify things, Biarri is offering its markdown tools as a managed service, so no IT work for you to do! We’ll do the analysis and provide you with the best possible markdowns.

Biarri’s tools are trusted by

Get in touch

  • Hidden
  • Hidden
  • Hidden
  • This field is for validation purposes and should be left unchanged.

COVID Cases in Australia

Optimised Health Care Rostering as a Managed Service

Coronavirus cases in Australia
The Coronavirus Status in Australia as of March 16 2020

With the outbreak of the COVID-19 pandemic, the most important aspect of our health care systems has been put to the test – people. It has resulted in longer work hours for all, with leave blocked out for the next foreseeable future and put an enormous strain on an already strained system – both emotionally and financially.

When things are difficult like this, the best we can do is to leverage the best tools possible to get ahead of these situations. And it is not just N95 masks and ventilators that hospitals need to function properly. Even more important than this are good rosters.

The Importance of Good Health Care Rosters

As helpful as ventilators and N95 masks are, without great people around to use them, they are no use. However, having great people is only half the story. The other half is giving them the right amount of work to do so they don’t burn out or work too long hours. It has been shown that poorly rostered staff who:

  • Work at least 12 hours per day are associated with a 37% increased hazard rate and
  • Work at least 60 hours per week are associated with a 23% increased hazard rate

When it comes to Coronavirus, the numbers become even more frightening as the effects are worse than simply longer hours. A new study, published in the Journal of the American Medical Association, shows just how bad it can get.

The survey-based study examines the mental health outcomes of 1,257 health care workers attending to COVID-19 patients in 34 hospitals in China. The results put the negative effects of overworking in perspective. Significant numbers of people reported experiencing symptoms of depression (50 percent), anxiety (45 percent), insomnia (34 percent), and psychological distress (71.5 percent).

These statistics are threatening the health of not only our medical staff but also patients and so it is essential that we start developing better rosters immediately to reduce the overworking and also the mental burden.

However, there is a final issue just as important as the above which can also end up costing lives but is much easier to solve. The issue is simply about who does the rostering.

Biarri’s rostering tools are trusted by the following health care providers

Every Person Counts – Rostering Medical Staff

Although it may seem strange, in a typical emergency department it is often one of the consultants themselves who writes the rosters for the other consultants. In the time of a pandemic, we need every skilled person working and not having them spends days writing rosters for their colleagues.

In Biarri’s rostering work across many health departments, we have calculated that rostering an emergency department consumes around 20 days of clinician time per annum for each roster. In many emergency departments alone they can have four or five rosters covering:

  • interns,
  • residents,
  • registrars,
  • fellows and
  • consultants.

All these rosters add up to a lot of wasted days and in the times of COVID-19, by using optimised and automated rosters, we could save dozens of lives and provide more essential care to people who need it.

Optimised Health Care Rosters as a Managed Service

Based on the above benefits and the imperative of having all qualified doctors best utilising their skills, Biarri is offering its automated and optimised roster creation tools as a managed service. Out of demand from numerous hospitals and the need to provide the service quickly without any IT work whatsoever, Biarri is now offering hospitals access to its automated, dynamic and optimised roster generators for the next 6 months as a managed service. During April 2020, instead of acquiring a new tool and having to learn how to use it, Biarri will carry out this work on the behalf of the hospitals.

If this is of interest, please send an email to and we’ll get back to you immediately as to how we can help manage the creation of optimised and compliant rosters.

The Benefits of Optimised Rosters

The benefits of automated and optimised rosters are many, and we have detailed them elsewhere, but the most important features and benefits of Biarri’s optimised, dynamic rosters are:

  • Customisable to capture working rules and payroll requirements.
  • Make rosters more evenly distributed from a fatigue perspective (reduce burnout)
  • Rosters account for staff requests (RDO’s leave, availability times & specific shift requests).
  • Creates equitable rosters with required training and appropriate mixture of shift types.
  • Ensures the right mix of skills is on the floor always
  • Can generate preferred shift patterns with night shift blocks & weekend, day off grouping.
  • Minimum break rules based on type and time of shifts.
  • Demand driven shift generation rostering only those staff required.
  • In a health care setting, it typically takes 2 senior doctors up to 3 days to create a 10 week roster. With an automated and optimised rostering system, it can be done in under 30 minutes.

So reach out to us now to get your team focusing on what they were trained to do and not struggling to produce sub-optimal rosters leading to possible staff burn out, depression and anxiety issues.

Get in touch

  • Hidden
  • Hidden
  • Hidden
  • This field is for validation purposes and should be left unchanged.

AI Driven Digital Transformation

Part 2

Part 1 is here.

Finding an AI Suitable Business Problem

About a year ago, I began working on one of my most difficult projects for the Victorian Government. It was a project that quickly looked like it would end in a consulting train wreck and possibly cost me my career.

The project began with our team sitting around a desk with the client and being handed multiple data sets. We were then told to “do some AI with it”. Not only that but we were given four weeks to produce something of value that wasn’t obvious to the client.

Given the time frames, we got straight into the task but approached it a bit differently given we knew we were missing a key part of the project – the business problem.

The first two weeks passed quickly with us tackling both cleaning and preparing the data but most importantly, one of us was spending a significant amount of time with the client trying to understand their current challenges and link these challenges back to their data and then request more data as we realised what was missing.

By taking an agile approach and having a highly collaborative team (with the design thinking part working closely with the analytics part of our team), we were able to navigate a challenging project and “do some AI” with the data set.

We discovered a business problem which we could share valuable insights on and quantify people’s intuitions, which helped the client as it killed the HiPPO[1] and replaced it with facts.

Quantifying gut feelings then led to several clear actions for the client to take which improved their service delivery, i.e. their core value proposition. We will see below why this is important and how it can determine the difference between failure and success of AI driven digital transformations.

Although not recommended, this approach to analytics succeeded but how should you approach such a challenge if given the choice?

Via the Business Problem

Someone once asked me,

“How do I understand the value of a data transformation? I mean, how can I know what it’s worth?”

The answer to this question is simple – on its own, a data transformation is worth nothing.

To be valuable, any data driven activity needs to solve a business problem. That is, why are we joining this data or building a massive data warehouse? What is the point of this analysis?

Most importantly, what action will be taken based on this analysis?

These questions lead us to the answers of the final questions we had in the previous post.

This idea of analytics leading to action is so important that at Bunnings the internal analytics team has three tenets to their analytics work with the third part being the most important:

  1. Aggregate
  2. Analyse
  3. Act

More than just creating a data warehouse, the purpose of analytics work is to derive insights and then do something based on that. Bunnings may have refined its approach over years but what should you do now?

The Right Way to do Analytics

There are many recommendations as to how to approach the challenge of deriving value from analytics. One was proposed by Gartner and looked at a company’s analytics maturity – however, this was wrong.

According to Gartner, a company’s AI capabilities are measured on 4 levels:

With each of the levels answering the following question:

  1. Descriptive: What happened?
  2. Diagnostic: Why did it happen?
  3. Predictive: What will happen?
  4. Prescriptive: What should I do?

As we saw in the Victorian Government example above, what counts is to have a business problem to solve and then to find a more efficient, digital way to solve this problem – not analytics maturity for the sake of analytics maturity. Once we have the business problem, and if we are new to the analytics space, we can use the Gartner model not to go too deep too fast and risk derailing the effort.

Finding the Right Business Problem

So we now agree that we need a valuable business problem to solve. Best is a corporate challenge burning across a P&L or balance sheet right now. Problems people currently face are ones they want solve – now.

However, how do we find the ones suitable to ignite an AI driven digital transformation?

The suitable challenges can be found at the core of the business. The challenges conducive to leading to something big are ones that form part of the chain of delivery of a business’ key services – the challenge itself doesn’t need to be big, it can be small but still have a big outcome!

For example, an amazing Australian success story, which successfully carried out an AI driven digital transformation, was by a company once called Aerocare.

Aerocare was founded in 1992 to provide ground services to airlines and airports. If you’ve ever looked out a window after landing at an Australian airport and see a team unloading your baggage from the aircraft, you are most likely looking at former Aerocare employees.

Aerocare’s value proposition is the cost-efficient delivery of people and equipment to quickly and professionally load and unload aircraft. Their value chain begins with forecasting airline service requirements at airports and ends with the unloading and then reloading of the plane on the day and time of the arrival of the plane (even if different to the plan).

If we examine the value chain and look at where an AI driven digital transformation can have the greatest impact, it is anything that improves Aerocare’s delivery of services. Aerocare also understood this, so they quickly identified a bottleneck in the Excel based rostering of people to deliver their services at gates and planes.

As the number of Excel tabs began pilling up and choking calculations – leading to weeklong rostering activities, the AI driven digital transformation opportunity had arisen. No unnecessary data warehouses for the sake of it, whatever was to be done would lead to change.

So Aerocare realised they needed to replace the current Excel based processes with AI tools to automate the roster creation and delivery.

This meant little or no change management – just better rosters. People followed their previous processes but with better tools. Like our health client, we’d found a part of their service delivery which could be easily replaced with AI tools. This then fundamentally transformed the way they delivered their services without changing the way the company did business as well as avoiding the risks associated with a large digital transformation.

Very quickly this led to more efficient allocation of resources, more competitive bids for service provision at airports and significant company growth. The result for Aerocare?

In 2017, the global ground handling giant Swissport acquired Aerocare for an undisclosed sum of money.

Why Did This Approach Work?

In 2015, MIT Sloan review published an article which incidentally revealed why the Aerocare approach to AI driven digital transformation worked. They stated:

“[M]aturing digital businesses are focused on integrating digital technologies, such as social, mobile, analytics and cloud, in the service of transforming how their businesses work. Less-mature digital businesses are focused on solving discrete business problems with individual digital technologies.”

The Aerocare solution seemed to solve a discrete business problem, however, it was one that transformed how the business worked, as intended in the first part of the MIT Sloan quote.

So we don’t have to look to uproot an entire organisation – we just need to find that part of the service delivery that is currently being done on a regular basis with:

1. Paper based tables; or

2. Excel sheets; or

3. Complicated, manual database queries.

Assume some general delivery of services or products as shown in the following diagram:

Then if any of the bullet points in the white boxes are being carried out using one of the above three manual methods then each could be tackled independently and easily via an AI tool. Once we’ve found the part of the service delivery in the above stylised diagram which is like 1, 2 or 3 above, then it is time for AI driven digital transformation.

Turning Data into Value – AI driven digital transformation

The answer to the success of the transformation question seems simple but it is only so in hindsight:

Find the key part of your service delivery that is not automated. This can most likely be recognized by the usage of Excel or regular, manual database queries.

Then replace this part of the service delivery with AI tools. The value is clear and immediate so let’s make clear what we’re doing:

Replace error prone, slow, manual and insecure processes with robust, fast, automated and secure AI techniques.

This now can resolve our question we posed earlier:

“How do I understand the value of a data transformation? I mean, how can I know what it’s worth?”

By choosing a business problem associated with your service delivery, it then becomes clear and easy to identify the value of the improvement as the change will directly improve the bottom line. It will improve the way your business operates and the value of the business. AI driven digital transformations allow one thing above all – more profitable and faster scalability via better decisions and automation.

Not only does this show us how we turn the siloed piles of data into value (we don’t touch them unless necessary), a McKinsey study showed that digital transformations that use some form of artificial intelligence, are almost 50% more likely to be successful. AI is one of the greatest factors that differentiates companies with successful digital transformations from those that aren’t.

So where does your Excel bottleneck exist? What part of your business can you transform with an AI driven digital transformation?

Reach out to us below to find out how and discover more companies embracing AI.

[1] Highest Paid Person’s Opinion

Get in touch

  • Hidden
  • Hidden
  • Hidden
  • This field is for validation purposes and should be left unchanged.

AI driven digital transformation

Your AI Driven Digital Transformation

Part 1

Even in the world’s best operating theatres, patient outcomes depend on more than the surgeon or their equipment.

Currently in hospitals around Australia, when and even whether you see a surgeon isn’t due to a medical expert but a booking clerk. To understand how this works, the following diagram outlines the current process in most hospitals:

After a hospital receives and approves your Request for Admission, to book you into an operating theatre, a booking clerk will often begin by recording your potential booking date in a notepad or Excel sheet. Only after reviewing hundreds of other patients will the clerk finally secure you a spot in the hospital software booking tool, once they are confident the proposed time is the best outcome available – that they can find.

This intractable game played by booking clerks is a gigantic puzzle placing thousands of patients with varying operation lengths and medical needs into a limited number of booking slots. Simultaneously the clerk needs to pay attention to the severity of the condition, the maximum permitted length of time until the guaranteed operation, other pre-operative complications and much more. It is a difficult problem with a range of constraints and variables making it very difficult for even the best booking clerk to optimise. 

It is also one of the greatest success stories of AI driven digital transformation.

Improving patient outcomes does not require complex change management, more human resources or redesigning the extant booking processes – just let AI take care of the ordering and optimisation of which surgery slots are given to which patients. This relieves experienced medical and clerical staff of the laborious and complex allocation process and leaves them free to focus on the quality control functions of reviewing, amending and approving an already optimised schedule.  

This type of improvement is a welcome one, as patient waitlists across Australian hospitals remain a source of patient dissatisfaction and are only set to get worse with an aging population.

So what can be achieved with such a seemingly straight forward upgrade to a booking process?  In the current health care system, even a 1-2% improvement would be revolutionary.

Biarri has been fortunate to have recently rolled out such a tool at a large health provider in Australia and in this case, the results so far have shown double digit improvements in operating theatre utilisation – double digits.

This is a step change improvement that shows the true potential of intelligently integrated AI tools. We estimate that it will lead to a reduction in the cost to treat a patient by around $200 / patient. With hundreds of thousands of patients going through the Australian medical system alone, this will save tens of millions of dollars.

What this AI driven digital transformation shows is that we can reduce our waitlist faster than ever, with the same number of resources. It is the gigantic puzzle of allocating people to surgical slots that is causing people to miss lifesaving surgery and it is a task perfectly suited to AI – and it exists in all industries.

In this set of articles, we’ll discover how to achieve this type of transformation in any organisation. This first article sets the scene and the second article will reveal how to create the value.

But before we discuss AI driven digital transformations, what is a digital transformation in the first place? And what change does AI make to it?

Digital Transformations

Digital transformations are about using digital technologies to create new or modify existing business operations, e.g. the day-to-day processes to provide a service, customer experience or a transaction.

When people speak about it, a digital transformation is often used in the sense of replacing one simple digital technology such as Excel with a large scale, holistic software system (often cloud based), especially in an enterprise with thousands or tens of thousands of employees.

The first digital transformation to typically occur in an enterprise is the implementation of a back office tool to manage finances (general ledger managing the cash going in and out, receivables, invoices, etc.), management accounting (budgeting, costing, etc.), procurement and more.

These capabilities are often rolled up into what is called an Enterprise Resource Planning (ERP) tool. Other major digital transformations, which often happen separately to ERPs but can be a part of it, are implementations of CRM tools which go beyond just the storing of customer data and create triggers and actions based on the characteristics of the customers’ behaviours – a first step on the journey to AI.

And the AI part?

Artificial Intelligence in industry typically refers to a set of mathematical techniques that allow one to make predictions or prescribe what actions to take. In the modern sense it often involves lots of data and applying analytical techniques to this data to answer questions like:

  • How many sales can I expect tomorrow?
  • Where should I best locate my products to meet this demand?
  • How can I route my vehicles to make sure those products are in those locations in time?

AI and digital transformations should go hand in hand because whatever the digital transformation looks like, it always leads to one thing – lots of data. So what can we do with this data to create value via AI?

Data – A Curse or a Boon?

Let’s shift gears for a moment and assume you now have begun your digital transformation and the data is flowing. You look at your data sets and a sinking feeling sets in.

After carefully gathering requirements, making sure you used the right mix of cloud technologies and then digitising the processes with the right tools, you realise your next big challenge:

The data from each system sits in its own silo.

As the complexity of this task sweeps over you, your boss bursts through your door, breaking your train of thought, and barks at you:

You’ve got all this data, now what? How do you start creating value? And I’d like to know by the end of today.

It dawns on you that this is the whole promise of the digital transformation that you bought into – a holistic view of your operations. Given what your boss has now demanded of you, how can you create value quickly?

Do we need to combine all this data now?

Is a large, expensive data warehouse the first step to the answer?

By choosing the right analytics battle, as we saw with our health client that the solution to the previous questions can be quite simple and deliver enormous benefits but many stumbling stones lie ahead of you – the first of them being your migration to the cloud.

Data Lake

Creating a Lack of Value

The trend to migrate to the cloud is currently one of the biggest tech trends of the last three years. With over 85% of enterprises now having a multi-cloud strategy, there seems to be no stopping its inexorable steamrolling of anyone expressing a different opinion.

However, as cloud migrations are creating a crowded sky, more than one in three cloud migrations are failing – even when we try and measure it against any metrics of your liking such as:

  • Reducing costs;
  • Improved staff productivity; or
  • Increased revenue

At the same time, around forty percent of global information workers are circumventing IT policies to try and maintain their productivity. The icing on the computational cake is that the productivity of employees is experiencing a gradual decline due to the complexity of current IT strategies and policies. So as you add your new layer of “cloud efficiency” on top of the existing IT systems, you recognise there is a risk that things may not turn out as planned.

So where is this taking your digital transformation agenda?

A 2017 Dynatrace study revealed that:

IT complexity and performance challenges are killing digital transformation initiatives, and causing organizations significant digital performance problems as often as once every five days.

The digital transformation to the cloud may be harder than people anticipate but what about the analytics/BI part of the transformation? Surely once you’ve sorted out the digital transformation challenges, the rest flows from there, right?

Not always.

A Logicalis survey recently showed that typically 60% of CIOs rate their organization as 3 / 5 or less when it comes to deriving value from BI and analytics work.

In fact, those who have derived value are few and far between and many would like to do something about that: less than 20% of CIOs say their organisations are using AI. Like your company, they’ve probably built the data warehouse in the cloud but they are still yet to derive insights and take actions from it.

What is going wrong?

With 66% of CIOs wanting to see AI adopted in their organization in the next 3 years, it is not a lack of will. Digital transformations have been praised as the next industrial revolution that is supposed to lead to untold benefits and chart busting efficiency gains. So where are they?

In the next part we will find the surprisingly simple way to achieve the promised gains without reinventing your business. By focusing on an AI driven digital transformation, you will increase the chances of success of your digital transformation and quickly show some early wins.

Get in touch

  • Hidden
  • Hidden
  • Hidden
  • This field is for validation purposes and should be left unchanged.

Better Digital Experiences

The State of AI in Marketing

As Artificial Intelligence (AI) permeates almost every part of every business, it is important to stop and review how AI can help you do your job better.

One field that has been dramatically affected is marketing and so it was timely that WP Engine has now brought out a survey with Smoothmedia and Vanson Bourne to reveal how the marketing landscape is being disrupted by AI.

The Results

As part of the panel of experts interviewed for the report, Biarri has been provided exclusive access to the report here.

Biarri’s key contribution to this important research was revealing the way data is best used and managed to deliver practical results to marketing departments without setting false expectations and over promising things that AI cannot deliver.

Download the report to discover what trends are changing the way marketers exploit AI to achieve results such as:

  • 42.5% of businesses seeing a visible increase in sales with AI
  • 37.5% increase in customer satisfaction
  • 29% increase in website visitors

Positive results such as these have led to around 32% of businesses planning to increase their AI budgets by more than 50% over the coming period. So we expect to see a separation in the market of those businesses which pull ahead of their competitors with AI and those who see their market share eaten away by AI driven experiences.

The key is using AI to deliver better digital experiences which leads to more relevant and personalised content. However, a question which many marketers must ask themselves is will this personal data be used for unethical purposes? The answer from the research was clear: 57.6% of survey respondents said they believe AI will ultimately have a positive impact on the world.

Biarri and AI in Marketing

Biarri’s team has a depth of experience in delivering retail and marketing AI solutions around:

  • optimal channel management,
  • market mix modelling,
  • pricing optimisation,
  • life-time value modelling,
  • anomalous behaviour,
  • churn modelling

and much more. Reach out to us now to find out how you can leverage AI to improve your marketing performance. We have conducted our own analysis to be able to show you the pathways to fast value with AI tools.

Get in touch

  • Hidden
  • Hidden
  • Hidden
  • This field is for validation purposes and should be left unchanged.