Acquisitions large and small bring with them a laundry list of post-merger system and process integration challenges. Consider the leading global education company Houghton Mifflin Harcourt (HMH). When Houghton Mifflin acquired fellow education provider Harcourt in 2007 for $4 billion and migrated Harcourt’s data onto its existing SAP ERP system, the finance group at the newly combined entity struggled to aggregate the data it needed to produce timely revenue forecasts. And these were not trivial forecasts; they drove many of the company’s core supply chain decisions, including how much inventory to buy and carry, where to locate that inventory, and how much it might cost.
But before the finance team at HMH began searching for the right technology to address the challenge, it first undertook a thorough process-mapping project to understand where the data was coming from. Not only did that project uncover redundancies in the company’s operations, but it also helped HMH implement a solution that would cut its revenue forecast cycle time from six weeks to several days.
The HMH of today was created through a history of mergers and acquisitions. Its largest business unit is the Education Group, an intricate mix of products and sales channels that accounts for 95% of the company’s revenue. Such a complex business requires a number of data inputs to create an accurate revenue forecast, and involves a number of business units eagerly awaiting their own versions of the forecast to plan their operations. Hazel Hughes, Senior Vice President of Finance and Business Control at HMH, says that creating and distributing revenue forecasts after the merger required a lot of effort and extreme accuracy to avoid input errors.
“There was a lot of sheer manpower required to get the work done, especially on the revenue planning side,” Hughes recalls.
Much of that forecasting work fell on Mary Jane Rivera, Vice President of Finance, who is responsible for the revenue side of the business. To create the revenue forecast reports, Rivera and her team had to pull year-to-date actual sales from SAP NetWeaver Business Warehouse (SAP NetWeaver BW) and then pull data on open orders before going into Salesforce.com to review what opportunities were won and lost. “Then through the depository group, we’d find out how many sales were not in the system yet and tap into the legacy systems to get any historical data,” says Rivera.
(L to R): Bob Verdin, Finance Manager; Mary Jane Rivera, Vice President of Finance; Kenneth Poage, Finance Manager; and Steve Nikolovski, Senior Financial Systems Analyst
Following this method, the only way to analyze the information from these disparate systems was to port large amounts of data into Microsoft Access and Microsoft Excel, which was an inefficient and error-prone process. “We knew we weren’t where we should be in terms of our business intelligence and our technology roadmap,” Rivera says. “When you’re trying to drive forecasts and reporting that provide any level of visibility, you have to be able to get down into the details. And that was very difficult for us in that environment.”
Process Mapping Shows the Way
According to Rivera, HMH knew the root of the problem was a lack of good tools. “But, we wanted to review our processes first to ensure our technology map fit our process map,” she says.
As a business formed by mergers, HMH’s processes were a mixture of different companies’ practices. So the finance team decided to undertake a detailed mapping of the current sales and operations processes to understand which groups should be providing what data to create the forecasts — and how the forecast reports should look when distributed to the various stakeholders.
“We reviewed our entire organization’s processes so we knew what all of the groups were doing and where the handoffs should be,” Rivera says. “It required a lot of patience to complete that process mapping because it involved working with many groups that had competing priorities during a time of change. We had to manage those competing priorities and derive a solution that was optimal for the entire organization, not just for those individual groups.”
Reviewing processes and roles in the company with an eye toward financial planning uncovered some redundancies (and data inputs) that resulted from the merger. Identifying those areas not only helped the finance team in its quest for faster data, but also helped move the entire company forward in its reorganization.
“We discovered that various groups were doing the same thing in slightly different ways, such as defining a data element differently,” says Rivera. “We began to understand the purpose of those processes and found ways to complete them sooner upstream and resolve the downstream redundancies. Without doing a full-blown process mapping exercise, we would’ve missed some opportunities for streamlining.”
(L to R): Richard Vosilla, Vice President of Finance; Hazel Hughes, Senior Vice President of Finance and Business Control; and Jennifer Buko, Director of Business Intelligence for Finance
The process mapping also confirmed to the finance team that the handoff of the revenue forecast to the supply chain team was occurring far too late to be effective. Having outdated information doesn’t allow the supply chain team to purchase and place inventory effectively to meet expected demand, leading to missed sales opportunities and costs incurred from obsolete products.
According to Hughes, another benefit of the process mapping project was that it presented a clear business case for a new solution in the finance organization by showing how compressing the revenue forecasting cycle could be profitable. “Because we were using outdated forecasting processes and data, we were spending a lot of money creating potentially obsolete products,” she explains. “The business imperative around the revenue forecasting cycle time compression had such a high payback — that piece alone made the business case for the entire implementation.”
A Game-Changing Solution
“We just continued to tweak and refine the data until we were happy with how it looked inside of SAP BusinessObjects Planning and Consolidation.”
— Mike Onderdonk, Senior Vice President of Business Intelligence, IT, HMH
“With a sensible and effective process design in place, we then set out to find the technology we needed,” says Rivera. The goal was to find a solution that could aggregate data from a variety of independent sources and produce an accurate and timely revenue forecast.
While the IT organization certainly helped facilitate the project, Jennifer Buko, Director of Business Intelligence on the finance team, says, “This was a business-driven implementation, not an IT-driven implementation that was overly-dependent on IT. We were very passionate about the end product so the finance organization put its time and resources into this project. All of the development and support of this is housed within finance, although IT was a key business partner throughout the implementation.”
Mike Onderdonk, Senior Director of Business Intelligence on the IT team, says that the top priority on the IT side was to find a solution that runs on standard hardware and operating systems. “Since this product would require interfaces with various SAP and non-SAP data sources, we wanted to be sure it was compatible with our SAP NetWeaver BW environment — SAP BusinessObjects Planning and Consolidation was the tool that fit best with our long-term technology footprint,” he says.
Onderdonk says the finance team’s process mapping made the implementation much easier. “When it came time for the interface work, we knew what elements users would be pulling into the application,” he says. “So we could jump right into the actual mapping of data fields from SAP NetWeaver BW to SAP BusinessObjects Planning and Consolidation. That saved us a lot of time.”
Another big time-saver was creating sample queries out of SAP NetWeaver BW to ensure the data elements were correct before IT wrote the actual interfaces. “If we coded these hard interfaces beforehand, and the data elements didn’t work, it would have taken us a lot longer to make changes,” Onderdonk says. The planning and consolidation application’s ability to upload data from a file also streamlined the systems design. “We just continued to tweak and refine the data until we were happy with how it looked inside of SAP BusinessObjects Planning and Consolidation,” he says.
“This was a business-driven implementation, not an IT-driven implementation that was overly-dependent on IT.”
— Jennifer Buko, Director of Business Intelligence, Finance, HMH
Today, the application receives daily feeds from both SAP and non-SAP sources, including the number of shipments made, a direct feed from Salesforce.com to see opportunity wins and losses, and a direct line on open orders and even depository sales. As a result, the finance team can now produce revenue forecasts in several days instead of several weeks, providing the operations team with timely, actionable data on where sales will be coming from in the forecast period.
Putting Data into Action
This timely forecast data is driving operational improvements across the organization. The operations and supply chain groups are receiving customized forecasts in a much cleaner format than they did in the previous spreadsheets because of the logic built into the planning and consolidation application. The global supply chain group now gets a flat file that breaks the forecast data down by state, program, and grade level. “We now can see what programs will generate revenue within the forecast period, so we can build out improved purchasing plans,” says Rivera.
Knowing where sales will come from also lets the supply chain group put the right inventory in the right location at the right time. The SAP BusinessObjects Planning and Consolidation rollout was one of several optimization initiatives underway at HMH that produced materially improved results for the company and its customers.
“The business imperative around the revenue forecasting cycle time compression had such a high payback — that piece was the whole business case for the entire implementation.”
— Hazel Hughes, Senior Vice President, Finance, HMH
According to Hughes, customer fill rates jumped from 70% to 85% — while on-hand inventory dropped 8% in a year in which total sales were up by 10%. Incoming calls from customers also dropped by 23% in the same period.
HMH compressed its revenue forecast cycle by one month and improved its forecast accuracy by close to 15%, driving down excess inventory. And that small improvement reduced the cost of carrying excess inventory enough to cover the cost of the implementation in roughly two months. “The implementation heavily contributed to avoiding tens of millions of dollars of untimely and, in many cases, unneeded purchases,” says Hughes. “Certainly, the improved processes and technology resulted in savings of millions for reduced product obsolescence, as well — almost 1% of sales.”
Beyond the supply chain organization, the sales operations group uses the more timely revenue forecasts to develop sales quotas. And finance and other corporate groups use the forecasts to gain a broader view of the business, determine what business areas are expected to grow or decline, and build strategies for those trends.
“We developed a number of automated reports that provide revenue by various slices,” says Rivera. “And we have very industry-specific reports, like adoption calendars, which are very time-consuming to create, so we already built the logic into the application to do that.”
With revenue forecasts much more accurate and timely, the cost side of the finance team can also make much more accurate projections on what the company will spend. “Everything we do is driven off those revenue projections that provide the volume inputs into our cost assumptions,” explains Richard Vosilla, Vice President of Finance, who is responsible for the cost side of the business. His team builds cost forecasts based on various cost drivers, such as paper, printing and binding, transportation, and even royalty costs — all of which are based on sales volumes. Sell more books, and the costs go up. So the improved revenue forecasts and the accessibility to SAP BusinessObjects Planning and Consolidation are opening up new possibilities.
“One of our biggest challenges is providing several views of our data to the groups that need it. This application lets us enter in the underlying data once and then spin the data however we need to for our stakeholders.”
— Richard Vosilla, Vice President of Finance, HMH
Currently, Vosilla’s team is building models to identify and capture those critical cost drivers to allow for detailed forecasting on the cost side. Onderdonk says the work done in the revenue phase will streamline the IT side of that project. “Because the process is similar to what we did on the revenue side, we already nailed down the approach, and we’re repeating much of the same process. It’s just different data that we’re sending.”
Vosilla adds, “One of our biggest challenges is providing several views of our data to the groups that need it. This application lets us enter in the underlying data once and then spin the data however we need to for our stakeholders.”
With roughly 25 users of SAP BusinessObjects Planning and Consolidation today, HMH plans to expand access to more users at various levels. “We expect our use of the application will have continued phases of improvement and evolution in our business planning environment,” Hughes says.
Both sides of the finance team — cost and revenue — agree that the application allows them to perform more customization within the finance organization without requiring IT’s help to set passwords. And beyond the finance team, HMH’s global supply chain organization is evaluating the possibility of using the application for more specific demand forecasting reports.
With much of the data flow mechanics now automated, Hughes is already thinking about moving into more advanced analytics capabilities, such as exception reports and alerts that indicate when a reforecast is needed. “The system could tell us when, for example, an expectation is trending off by more than X% and we need to reforecast that specific element versus the entire detail,” she says. “Having this fundamental stage in place will allow us to move to that next level, such as tracking and adjusting for field input biases.”