For budgeting, planning, financial reporting, and performance analysis, spreadsheets are clearly inefficient. Yet the alternative is not always clear-cut. Finance managers who have successfully used software to streamline their business performance management (BPM) processes and free up their team's time for decision support almost unanimously loathe the idea of returning to manual, spreadsheet-based budgets. But many companies -- even surprisingly large ones -- still use Excel extensively, if not exclusively. Why?

Most market analysts believe that one contributing factor is the frequency with which other major software implementations have run out of control. BPM systems may invoke memories of the nightmare ERP deployments of the '90s. Implementation costs can be hard to anticipate in advance because how much customization a software product will need and how much effort data integration will necessitate aren't always clear after the sales demo.

That's why BPM Magazine set out last fall to quantify how much time and money companies tend to spend on BPM software implementation projects. Clearly, every initiative is different and our survey results can't tell an individual organization what will be required to meet its unique needs. But the information we gathered from more than 200 readers provides insight into projects' common levels of resource consumption -- and provides some interesting insights into what project characteristics make for happy software implementers.

The Fundamentals

All in all, the news from our survey is positive. Companies tend to select software that aligns closely with their needs, and they tend to be pleased with the fruits of their BPM initiative. Twenty percent of survey respondents said their software implementation process involved no customization of the application; they changed their processes to fit the software, rather than changing the software to fit their processes. Fifty-six percent customized the software a little bit but chose the product carefully to minimize the amount of effort their implementation would require. Only 23 percent reported that their initiative involved a lot of customization to the software.

Few companies are taking full advantage of all of their software's capabilities, but most are getting a lot of mileage out of the application. Four percent are using less than a quarter of their product's total functionality, and all of those expected before implementation to use more, but they are a tiny minority. Fifteen percent of survey respondents reported that they're using more than 90 percent of the software's total functional capabilities, and 36 percent are using between 75 percent and 90 percent of their application's features.

To optimize use of the software, many companies undertook fairly costly and time-consuming projects (see What Does a BPM Initiative Really Entail?) Half of all respondents spent more than $100,000 on the software, and 4 percent spent more than $1 million. Fifty-five percent of budgeting and financial planning systems cost more than $100,000, as did 45 percent of operational planning and forecasting systems, 54 percent of external financial reporting applications, 35 percent of management reporting and dashboard products, and 60 percent of financial analytics systems. Most companies spent considerably less on consultants than on the software, but 5 percent spent more than $1 million on implementation consultants and 3 percent spent that much on help with the planning and software selection process. For 29 percent of survey respondents, implementation costs came in higher than expected, but for 18 percent, the system's setup actually cost less than plan.

The median length of a BPM software implementation project is five months. Over the course of those months, the project consumes a median 176 days of finance staff time, 88 days of the IT staff's time, and three days from staff within the company's business units. For 32 percent of respondents, the project consumed more finance staff time than expected, and 31 percent of projects used more IT staff resources than expected. However, half of respondents said finance staff time came in exactly on target, and 51 percent of projects met expectations for IT staff time.

Fifty-four percent of all survey respondents expected before implementation to use their software more fully. Still, most are satisfied with the outcome of the project. Fifty-five percent said the resulting system meets their expectations. Seventeen percent said it far exceeds, and 13 percent said it slightly exceeds, their goals. Twelve percent of implementations have failed to fully meet the software buyer's expectations, but only 2 percent of companies are very disappointed in their system. More important, the vast majority (86 percent) of respondents feel the BPM system they are now using is worth the cost and effort of the purchase and implementation process.

What factors affected implementers' satisfaction with the process? The most notable standout in our analysis is the degree to which companies use consultants -- but the effect may not be what you'd expect.

Consultants' Impact on Project Costs

Consultants necessarily add to the cost of any project, for the obvious reason that they must be paid. But companies may expect this spending to provide them with a better idea up front of what, exactly, the project will entail. Consultants should have the experience to lay out in advance how long the implementation will take, how much staff time it will consume, and how much it will cost. Those who hire these service providers -- particularly in the planning stage of the project -- should be the same companies that come in on time and on budget because they set their expectations based on expert guidance ... right? The survey results indicate that many readers are shaking their heads at that notion.

According to our research, implementation lasts about 50 percent longer for those who hire planning consultants than for those who don't. This is true at each level of spending on the software, and it's a trend that also emerges when companies are divided up according to their spending on implementation consultants. The median length of software rollout among companies that did not hire implementation consultants was three months; among companies that hired help with the rollout, the median length of the project was exactly twice that.

It might similarly seem that BPM software-implementation projects which rely heavily on third-party service providers would consume fewer staff resources; after all, the consultants should take on work that internal staff would otherwise have performed. However, in our survey, the opposite was very distinctly the case (see exhibit 2, below). Although the one survey respondent that paid more than $500,000 for software but hired no consulting help devoted vast staff resources to the project, the average project that involved planning consultants took up considerably more staff time than the average project in which planning and purchase were handled in house. The median level of finance staff time for an implementation in which the company hired planning consultants was 264 person-days; the median for companies that hired no planning consultants was half that: 132 days. IT teams saw a similar result. Companies that hired consultants during the planning/purchase process used a median of 132 days of IT staff time on the project; those that didn't used a median of 88 days. The accuracy of preimplementation staff-usage estimates wasn't improved by hiring consultants, either. Among organizations that did not hire planning consultants, only 2 percent invested substantially more finance staff time -- and 4 percent substantially more IT staff time -- in the project than they expected to. But among initiatives that involved over $500,000 in project-planning services, 50 percent took far more finance time, and 63 percent far more IT time, than plan.

Do implementations in which consultants are involved in the planning process also come in over budget? Again, the answer is counterintuitive, but the survey results are clear. Organizations that hire consultants during the planning and software purchase process are much more likely to underestimate the cost of the implementation project. Among survey respondents that relied on consulting help in planning the project, 38 percent spent more on the implementation than they expected to -- 17 percent went substantially over budget -- while only 3 percent spent substantially less on the implementation than they expected to. Conversely, among companies that did not hire consultants at all over the course of their software planning, purchase, and implementation, only 6 percent spent more than they planned to (just 2 percent substantially more), and a full 19 percent spent substantially less than they expected to (see exhibit 1 below).

At first glance, this data seems to suggest that BPM planning consultants routinely underestimate the length and cost of projects. Perhaps, instead, they see more opportunities for process improvements and software customization than would a company without consulting help -- and so encourage teams to take their implementation projects beyond their initial objectives, timelines, and budgets to accomplish gains that the internal implementation team didn't realize were possible. Of course, another explanation is that businesses tend to call in consultants primarily when they are undertaking a complex implementation, one that involves thorny data-integration problems and requires heavy-duty customization. In other words, the projects that consultants are hired to help with may be the same projects that would have run long, would have taken excessive staff resources, and would have come in over budget anyway.

This is a legitimate notion and one that likely played at least some role in our survey results. Unfortunately, it's also a theory that we can't test precisely based on our questionnaire. The one question we asked that may serve as an indicator of how much effort an implementation could be expected to take is how much customization the project entailed. One might expect that the projects for which companies required consulting help would be those in which the software was most customized for the buyer's specific needs. That doesn't seem to be the case.

Among companies that didn't hire consultants for either the planning/software purchase process or for the implementation (a quarter of all survey respondents), only 27 percent changed their processes to fit the software, rather than changing the software to fit their processes. Fifty-seven percent customized the software a little, and 16 percent said their implementation involved a lot of customization. Organizations that hired consultants did do more customization, but the difference wasn't extreme (23 percent customized a lot, 56 percent a little, and 20 percent didn't customize at all). Even more interesting, a full 86 percent of companies that performed a lot of customization without any help from consultants during either phase of the purchase and deployment process reported that the implementation's costs met their expectations. Among the few companies that spent more than $500,000 on consultants in both phases, only 43 percent said the cost of the implementation met their expectations. For 57 percent of these big spenders, implementation went substantially over budget.

So, Who Is Happiest With BPM?

Not too surprisingly, the larger the investment, the less likely finance managers are to be satisfied with the resulting system. Big spending and hefty consumption of staff time mean complicated projects and high hopes. More than half of companies in each of the survey's software-spending brackets said the resulting system meets their expectations. However, smaller projects are much more likely to exceed expectations than is software that cost over $500,000. And although only 10 percent of applications that cost less than $50,000 failed to fully live up to expectations, 44 percent of products that cost more than $500,000 are underachieving. Similarly, companies that are pleased with the results of their software implementation project used far fewer internal resources over the course of the project (see exhibit 5 below).

Another factor that appears to affect buyers' satisfaction with the software is the level of customization the project entails (see exhibit 3 below). Organizations that didn't attempt any customization of the software are more likely to have exceeded their goals. The difference becomes even more pronounced when the project's cost relative to budget is thrown into the mix. Among companies that customized their software a lot and came in over budget, just 16 percent of projects exceeded expectations (5 percent exceeded by far). Forty-four percent of these companies are somewhat disappointed in their system, and 11 percent are very disappointed. Conversely, among companies that customized their software a little but still came in under budget, 57 percent of projects exceeded expectations (43 percent did so by far). And just 4 percent of solutions for these organizations have failed to meet expectations.

Companies that hired consultants tend to be less satisfied than those that did not (again, see exhibit 1). Only 9 percent of those that hired consultants for help during the planning and software purchase -- and 11 percent of those that hired implementation consultants -- said their solution far exceeds their expectations, whereas 30 percent of projects that did not involve any external consultants far surpassed goals. In addition, among the handful of companies that spent more than $500,000 on planning-and-purchase assistance, half feel their implementation wasn't worth the money and effort it entailed (see exhibit 4 below). Whether the consultant contributed to the disappointment is impossible to determine, but the data indicates some kind of linkage. Still, it's important to keep in mind that even among organizations that hired consultants for both planning and implementation, less than 25 percent of all software systems have failed to fully meet expectations.

One factor that doesn't seem to affect companies' satisfaction is whether they install the software on internal servers or rely on an application service provider (ASP) to host it. A handful of survey respondents (6 percent) use BPM software that is hosted by a third-party provider. Although this practice is frequently promoted as an option for businesses that can't afford installed applications, most of the hosted solutions represented in our survey were relatively pricey. Thirty percent of respondents using an ASP spent less than $50,000 on the software itself, but 40 percent spent more than $100,000 on the software. Survey respondents that use hosted solutions spent about as much as the organizations that installed BPM systems in-house. Interestingly, too, satisfaction with the software -- and even whether the project came in under budget -- seems unaffected by whether the software is installed locally or hosted.

At every level of spending on the product, on consultants, and on implementation, some software purchases fail to live up to expectations. Most buyers who are displeased encounter surprises during the implementation process that substantially impact the speed and cost of the project. Unexpected data-integration problems were a common complaint among our survey respondents, as was a failure to anticipate the degree of customization the software would require. Some blame their software vendor for setting unrealistic timelines or delivering a poor-quality product. One noted that the licensing agreement contained hidden costs. Others blame their service provider; wrote one, "The implementation consultants didn't understand accounting."

But our survey also revealed that many BPM buyers are pleasantly surprised by the results of their rollout. Despite the risks of undertaking a project the size of the standard BPM implementation, few survey respondents regret having taken on the project. Many, in fact, said they would change nothing about their approach if they had the entire project to do over again. Implementation of a BPM software package is certainly not an initiative to undertake lightly. It will probably be expensive and eat up a good deal of time among finance and IT staff members, and possibly within the company's business units as well. But companies that enter the world of BPM with their eyes wide open are very likely to find the project well worth the effort.

Meg Waters is editor in chief of BPM magazine and of the monthly e-mail newsletter BPM Express.