HR Block: creating a competitive edge in a unique market

Demand forecasting has been an important and frequently utilized solution due to its ability to capture and combine historical trends with key factors to estimate future demand.

HR Block: creating a competitive edge in a unique market

For retail and many other industries, demand forecasting has been an important and frequently utilized solution due to its ability to capture and combine historical trends with key factors to estimate future demand. However, while forecasts are usually accurate when driven at a macro level, accuracy falls off rapidly as one drills down into specific segments.

Unfortunately, more actionable forecasts usually occur at a micro level. The ability to drill down to a more granular level and explain the nuanced demand in individual markets is key for district and store managers to understand where their business is succeeding and failing. Essentially, granular demand forecasts have the potential to be the most impactful, but are often the most challenging to model and are ripe for techniques like Machine Learning to solve.

H&R Block saw this as an opportunity to resolve longstanding gaps & gain a deeper understanding of what drives their business when it approached Neal Analytics with two goals: construct an advanced analytics solution to better forecast the total number of tax returns in the 2018 tax season, and gain a deeper understanding of the contributing factors driving that demand. On the surface, this appears to be a simple demand forecasting problem, but when you consider that H&R Block has the most seasonal demand in the S&P 500, or that traditional customer factors are not realistic predictors, this posed a unique and interesting challenge. Neal Analytics was eager to support where H&R Block had yet to succeed with other consulting firms or using internal teams.

Leveraging Industry Expertise

Given the limited timeline leading up to the 2018 tax season and the need to develop a sophisticated machine learning solution quickly, Neal Analytics targeted specific forecasting and business goals to quickly build an enterprise grade solution for critical milestones, leading to use in the 2018 tax season. By using Neal Analytics’ forecasting tools and leveraging H&R Block’s industry expertise, Neal Analytics provided H&R Block with accurate forecasts earlier in the tax season, as well as an indepth look at the key drivers of their complex business, thus enabling tactical adjustments to maximize performance in each company district.

Predictingan Evolving Market

The tax preparation industry is an established, yet due to culture shifts and technology innovation, constantly evolving ecosystem in which H&R Block has been a major player for over 60 years. With the growth of online preparation software, as well as continued competition from traditional competitors, H&R Block tasked Neal Analytics to provide them with key insights to increase annual return counts and market share. As an additional challenge in the industry, over 95% of their business is condensed to about 3 months, compounding the need for immediate intervention when things are off pace and requiring precise and immediate execution to reach seasonal targets. With the goal of arming H&R Block’s leadership with solid forecasts and insights on a week-by-week basis for the 2018 tax season, Neal Analytics hit the ground running, developing forecast models to predict the total number of tax returns to be filed in their offices. These models adjusted dynamically on a weekly basis based on actual returns filed the previous week, and Neal Analytics’ predictions for total returns were within 2% of the final number at the start of the season. Providing H&R Block with high accuracy so early on in the 2018 Tax Season gave their operations team more clarity and foresight on how best to plan office execution through the season, pivoting strategies on the fly as needed to capture maximum business. This solution armed H&R Block with new capabilities that would assist this established leader in staving off threats from their competition.

Office Level Forecasting

As a whole, H&R Block operates 12,000 offices that are broken up into a multitude of districts consisting of both franchise and corporate locations. Previously, return forecasts and strategic management were only run at a companywide level, with no concrete picture of what was driving variance in performance from one location to the next.

Neal Analytics aimed to resolve this gap by training thousands of individual machine learning models and utilizing the consensus predictions across various sensitivities for each market, providing reliable local forecasts and identifying the key drivers of increased returns.

Using the analysis of the key drivers of higher return counts, we were able to isolate the effects of price on demand and compare the effects of discounts to better understand trade-offs. By analyzing these effects at both the whole company and district levels, H&R Block was able to better understand in which markets certain pricing and discount strategies were the most effective. As a result, Neal Analytics not only provided accurate weekly forecasts for district managers, but also dissected the key performance drivers tailored to each manager’s small group of stores. One example of this relates to labor hours, by far the greatest expense in any services company. By utilizing their internal labor data, Neal Analytics was able to isolate the effects of adding or removing hours from a given district to see its effects on production. Ultimately, the success of the solution has ensured the final output will be integrated into H&R Block’s strategic planning for many tax seasons to come. Although H&R Block had previously attempted
to solve this problem with tools such as Tableau and open source code, they ultimately failed to significantly advance H&R Block’s analytics capabilities, leading to interest in what the Neal Analytics team could provide in thought leadership. H&R Block needed concrete ROI from a true business solution to prove the value from the cloud and data science. With this engagement, Neal Analytics delivered that ROI and the competitive advantage of Azure Machine Learning in only a few months’ timeline.

Transforming Data to Insights

Structuring returns data for the machine learning models also enabled a variety of BI dashboards. This context helped HRB gain insight into which regions are most efficient and productive, as well as provide clarity on how or where adding employee hours has the strongest impact on demand. To top it off, this analysis highlighted the districts with the lowest efficiency, arming operations with the tools to reel in employee hours and cut operational spend. Following this success, there is now an initiative to place Power BI analysis dashboards in the hands of sales leads, analysts, managers and executives companywide. This exposure to ML derived BI has generated momentum and innovation in a company with a promising enthusiasm for analytics and a desire to outrun its competitors.

Within this tax season, H&R Block was able to apply the analysis from the Power BI dashboards to identify the top and bottom performing districts and categorize key factors leading to their performance. These insights will become the main points of change in the 2019 tax season, but in the hands of district managers, also enabled tactical interventions and investments in the operations that most impacted their business.

As an example, H&R Block discovered that in a few districts, one of the most important factors in the loss of tax returns filed in stores is the retention of clients for more than 3 years. By highlighting this factor, H&R Block is able to pinpoint marketing investment and implement additional changes contributing to client satisfaction and thus improved retention. Another example of a key feature factoring into the loss of tax returns is the number of hours their tax preparers work in a week. Upon further analysis, H&R Block discovered that they may have been ramping down the number of hours assigned to a tax preparer too quickly, thus decreasing the number of returns filed per hour and increasing the wait times and client dissatisfaction. A district manager can now optimize their store’s operations so that they are investing in the right activities to maximize customer outcomes and increase annual returns.

On-Demand Store Analysis

Beyond an accurate number to report to shareholders and Wall Street, the primary value of the solution is derived from the ability to generate week-byweek analysis of performance. This enabled a pivot to better execution throughout the season instead of waiting until after the season is completed and making changes for the next. This ability has had a ripple effect of value to the business. Ed Dobbles, the Vice President of Analytics and Pricing at H&R Block, has attested to the impact of the solution saying “This solution fills a critical need for our business in that it provides my team and I with the data needed to go to the CEO with insights that result in action, positively affecting our clients and our bottom line. Further, we can deploy these Power BI dashboards to our store managers so they can see in real time how their stores are forecasted to perform, but most importantly, how to improve it.”

By addressing the demand forecasting and sales driver analysis problems in unison, H&R Block is now able to drive real time improvement to the management and operations of their corporate offices. Armed with this information from dynamic and clear reporting, their analysts and managers will better be able to handle next year’s operational changes and identify what changes are needed by implementing the best practices of top offices and intervening with their bottom performers.

The combination of Neal Analytics and the Microsoft Cloud came through for H&R Block, just in time for the tax season, allowing H&R Block to forecast earlier and more importantly, actionably interpret it, while setting up a long-term analytics platform in Azure with many additional use cases on the roadmap. We look forward to their continued success and consider this an outstanding customer outcome where exotic data science and AI were able to generate real business value in a rapid deployment methodology.

Future Development

With the 2018 Tax Season officially ended, H&R Block now has the capabilities to review their operation and execution of this year’s tax season, viewing performance through a few new lenses. By adjusting their overall strategy before the 2019 tax season, H&R Block intends to capture additional market share with better offerings, improve operations by focusing on what matters most to the customers in each market, and adjust proactively to the predicted events and market environment in the upcoming weeks.

Download