Houghton Mifflin Harcourt Publishing Co. is the world's largest publisher of educational materials for pre-K-12 schools. The Boston-based company is a relentless innovator, as its fresh-thinking approach to performance management demonstrates. BPM Magazine spoke with Donna Ritchie, HMH's vice president of financial services, publishing, and operations, about how performance management software is transforming the company's product development and purchasing processes.

BPM Magazine: You chose to focus your BPM project initially on the product development area. What were some of the challenges you faced there?

Ritchie: Product development spend in publishing is very complicated because our programs are so long-lasting; the development process can last anything from 18 to 24 months for a particular program. Some smaller things can be shorter — say, 6 months — but our projects often go as long as 24 months.

We have very large programs with numerous components. For example, a product might have various print components — student editions, teacher editions, assessments — as well as elements that we provide online. From the initial writing of a manuscript to the minute it goes to press, there are so many different activities. It's very complicated, with a lot of little contracts that make up very large amounts of spend.

Another complicating factor is the fact that we have a very fragmented vendor base. It would be much easier if we were, say, a GM and had just three big vendors to support us, but that's not the nature of development in our industry. We use a lot of individual freelancers, many mom-and-pop shops and small development houses, all the way through to big multinational production companies. It's just a very complex, data-heavy piece of the business to manage.

A third complicating factor is development time. With so many different pieces, keeping track of time — in terms of when we're going to spend against various portions of the program, what we've budgeted vs. reforecasting, getting the spend correct by month and by quarter — is very difficult.

These are some of the challenges that we attacked through this implementation.

BPM: Were you managing all of these processes through spreadsheets?

Ritchie: You got it! We used a very basic FileMaker database for commitments. HMH is a combined company, so obviously there were a lot of legacy systems in place. We also had lots of spreadsheets for budgeting and reporting, so we needed a system to assist with these areas as well. Just trying to compare budget with actuals — and to do so in a way that would provide insight into what was driving variances and controlling our various activities — was hugely time-intensive.

BPM: Did you spearhead the software selection process?

Ritchie: Yes, hand-in-hand with our IT team. We had a pretty robust process because we wanted to make sure that we were weighing IT's needs and ours appropriately.

One unusual requirement that was fairly high on the list was that it should be a Mac-compatible program, one that could be fed over the Web without a band-aid type of solution for Mac users. Other things on our list of decision criteria were price, database structure, customer service, and the vendor company's stability. We wanted to make sure that we had workflow functionality, and we needed a commitment to the purchase order portion of the project. We were concerned about runtime, latency, the application architecture, how well it linked up with SAP, and the quality of the implementation team. Those were our main criteria.

When we went through the evaluation, we felt that Tagetik hit each item that we cared most about. If there were bells and whistles to the other programs, they didn't stick out to us as things that we needed in order to get done what we needed to get done, and to do it in the way we optimally wanted to do it.

BPM: And as you started to roll out the system, you were faced with a reorganization …

Ritchie: Right. In fact, we almost tripled the size of the implementation. We restructured so that our development finance operations would cover all of the prior educational divisions, whereas it had started out as an elementary-only initiative. We started to cover elementary, secondary, and our supplemental products as well — a lot of different products, structured differently in terms of data and in terms of how they had been using SAP. The master data were different, and we're still cleaning up some of that.

Keep in mind as well that we hadn't ever fully integrated, from a systems perspective, after the Houghton Mifflin merger with Harcourt, although this was understood within the initial plan. But because we had these additional divisions, and none of these had been integrated either, there was definitely a multiplicative effect.

BPM: How did the BPM system help with the integration?

Ritchie: Tagetik replaced the spaghetti of legacy systems that were all kind of glued together. We had a very smooth integration with SAP. It's not a direct feed, it's a flat-file ETL [extract, transform, and load]. Tagetik was easily configured to enable that.

Now we have one place where we can go to input our budget. We have one standard template. We created a closed-loop system with a budgeting template that feeds into our vendor bids, which then sends feedback into our commitments module and purchase orders and links it all together with the actuals. We accomplished a lot by combining everything, putting it in one place, and having each piece of the process feed the next piece.

What we do in Tagetik is our budgeting, both at the top line and at a more detailed level, for product development investments. This is tied to our balance sheet accounts. We also manage our purchase orders in the system — commitments that are tied to the spend that's budgeted for our investments. And then, on the reporting side, we bring these two pieces together; the ETL link from SAP pulls the actuals back out.

So we're reporting on budget vs. committed, and spend vs. actuals, at a very detailed level: by program, over time, and by G/L. We are able to strip out variances in our spend that are driven by unit, by pages, and by rates (for example, where we decided to pay more for a particular page, for various reasons).

BPM: One of the first areas you focused on was the purchase order process, which is a bit unusual for the early stages of a BPM project …

Ritchie: Agreed.

BPM: What was the thinking there?

Ritchie: Our thinking was that if we were really going to have an effective budgeting and reporting system, then — given the length of our programs — we really had to have that insight into the middle of the process so that we understood how our commitments to our vendors were playing out, in terms of both the size and the pacing of that spend. For me, this is a critical piece of the puzzle; if you don't have that, you're missing the guts of the economics, which is all of the activity that you have going on that hasn't yet actually come back to you in the form of an invoice, but which you know you're going to be funding. Having that insight really helps with our planning.

BPM: I'm always surprised at how challenging it is for companies to improve their purchase order process. It often seems to be a real bottleneck.

Ritchie: Right. You want to get your purchase orders right, and you want to get them done in a timely way. We needed a system that could help us to drive efficiencies in the PO process itself because of the cyclical nature of our business. We tend to have very heavy periods of time, and we wanted to make sure that we're maximizing our efficiency so that we can absorb those heavy times more easily and generate less of a backlog.

BPM: What were some of the changes you made there?

Ritchie: Before we implemented Tagetik, there were a lot of different legacy systems in place. It was an antiquated process that, frankly, didn't have the right level of detail that we needed. For part of the process, we had a paper-based approval cycle. Now we have electronic approval for both Pos and invoices. We have an auto-bid loading functionality, so there is less hand-inputting of vendor bids, more automation in that process. There is still some manual work, though; there's no way to automate the whole thing.

BPM: What are some other benefits that you've seen so far?

Ritchie: We're still finishing up the implementation, but we definitely are already seeing benefits in terms of the validity of the data, the centralization of the data, and ensuring that we are linked to our core master data in the early days of the budgeting process. The purchase order piece is a nice, clean system. And it's easy for us to pull up the reports that we need.

BPM: How about on the planning side?

Ritchie: Tagetik's solution is helping us with longer-range planning by providing a central place to create a 2- to 3-year view of program spend and to keep that up-to-date and manage it as a living, breathing, overarching product investment plan. This is built into the budgeting function. We have two levels of budgeting: a high-level view — like a portfolio-planning or investment-level view — and a much more detailed line-item-level budget.

BPM: What will be your next move to extend the BPM implementation?

Ritchie: There are natural opportunities in the op-ex area. We're really just in the fact-finding stage, but I expect it to be a similar kind of situation: You have numerous different functional areas with detailed budget data that you have to build from and that eventually ends up in a divisional/corporate operating expense budget. No plans have been made at this point, but it's a natural extension for consideration.

INDUSTRY: Publishing

COMPANY SIZE: $77 million

System Setup

Product development budgeting, planning, and reporting: Tagetik

Enterprise resource planning: SAP