How to Understand the Basic Rules of Software TCO Calculation

Executive Summary

• Brightwork Research & Analysis has observations regarding software TCO.
• This article is an overview of how to leverage TCO.

Introduction

Total cost of ownership, or TCO, is the complete cost of owning something. TCO can be rearward looking—an accounting of what a purchase actually cost. However, in most cases, a TCO analysis results in a forward projection or forecast. TCO for enterprise software is the overall sum of the costs of the four main TCO categories:

1. Software Cost
2. Hardware Cost
3. Implementation Cost
4. Maintenance Cost

TCO is discussed in the abstract as a “good thing” but is rarely calculated in reality. If you think back upon all of the purchases you made throughout your life, how many of them included the TCO on the price tag, along with the actual purchase price? There is a good reason (or good reasons) for this and they go by the names of sales and marketing. The last thing a company wants is for their prospective customers to know the total cost of an item. The one exception to this rule is if the vendor has a study that shows their product or service has a lower TCO than that of a competitor’s product or service. TCO is the base value for ROI: it is the “I” in ROI. A TCO must be calculated before an ROI can be calculated. ROI is the formalized analysis of the universal ratio between costs and benefits, which I referred to previously, and is focused on both the revenue and the costs of an investment.

Determining the ROI of Enterprise Software

Determining the ROI of enterprise software is quite difficult because it means estimating the financial returns from software, which is a complex endeavor. Therefore, instead of producing ROI estimates, we apply a rating system to the application, which is used in conjunction with the TCO estimate. This still allows the value of the software to be estimated, but without the necessity for an ROI calculation. Accurate TCO analysis requires effort; getting into the real detail of the costs and benefits of applications requires a combination of first-hand implementation experience as well as the analytical ability to perform the analysis. This analysis should be built from a level of detail such that the TCO flows naturally from a more thorough analysis. While most software vendors hide their pricing information, all software vendors still like to discuss TCO in abstract terms as a “good thing.”

Item #1: Being in Favor of TCO?

Being in favor of TCO is like saying you are in favor of the American flag or apple pie; you’re in favor of “goodness.” You can’t actually find anyone who opposes TCO in principle. But the devil is in the details. As long as nothing is quantified or the TCO studies are rigged, all vendors can say they have the lowest TCO. Most software vendors are not only against publishing TCO estimations but are even against publishing pricing information for their applications on their websites. Software vendors cannot be relied upon to produce TCO estimates, and the best move is to disregard the software vendor’s estimate.

Item #2: Who Can Reliably Produce a TCO Calculation?

Obviously, TCO analysis should be produced by an independent source and not the entity that is selling the item. TCO is fundamentally a consumer concept, not a producer concept. Therefore the consumer or an entity with the consumer’s or purchaser’s interests at heart should perform the research and publish the results of the TCO. There is a research company that does this, but they serve the consumer market rather than the corporate market. This company is Consumer Reports. They perform TCO for consumer items. Consumer Reports is in a good position to do this because they take no advertising from any producer and only make their money from consumer subscriptions. The following is a TCO estimation for automobile ownership.

Consumer Reports calculations in seven common automotive categories show that the most expensive vehicle to run for five years is the Mercedes-Benz S550 at about $101,750. Consumer Reports calculated that the least expensive vehicle to run over five years was the Toyota Yaris with a manual transmission, at about $23,250.

One of the least expensive cars to own in our estimation is the small Honda Fit, which costs just over $5,300 a year to own for five years. It combines a relatively low purchase price with low depreciation, great fuel economy, excellent reliability, and fairly low maintenance and repair costs.

Paying more for a hybrid can save you money as long as you choose the right hybrid. Most mainstream hybrids that aren’t luxury or SUV models cost less to own over five years than their less expensive conventional counterparts. (Two exceptions are the Chevrolet Volt and Honda Insight. It takes six or seven years, respectively, to make up the added purchase price in fuel savings for those cars.)

Consumer Reports

These reports are valuable because they provide very actionable intelligence that consumers can use to make informed purchasing decisions and to avoid cars that have hidden maintenance costs. Consumer Reports is one of the few media entities that can say they are not swayed by vendor influence because they put a number of restrictions in place that are designed to counteract vendor influence. Consumer Reports goes to the extent of actually purchasing the products they review rather than relying upon free samples which is how almost all media entities operate that review products. This, of course, puts the media entity into the debt of the vendor, as well as allows the vendor to adjust the reviewed product so that it is perfect and different from the experience that the normal consumer would receive.

The Behavior of Consulting Companies and TCO

Consulting companies are constantly putting out falsified information about the cost-saving benefits of both ERP and outsourcing—which is not surprising as they are knee-deep in selling these services. Software vendors and consulting companies are not the only entities that seek to suppress TCO estimations, but IT analysts and buying companies often have little interest in this information, as well. Several entities disagree on whether or not a full TCO analysis should be a goal. In this section, we will review the concerns leveled frequently at TCO. One of the entities—which could be labeled as anti-TCO and is influential in the area of enterprise software decision-making—is Forrester, the IT analyst firm. In the field of enterprise software, one of the most amazing stories of the past several decades is the mass purchasing of ERP systems—purchases made without the customers searching for evidence that ERP systems are good investments.

If they had looked, they would have found that the logic presented to sell ERP systems had no evidence to support it. Costs are often described in general parlance as the amount that we pay for things. However, economists look at costs quite a bit differently. Promoters of ERP tend to present any benefits of ERP without acknowledging that the time and effort spent on the ERP project could have gone into other initiatives. The comparison should be between the gains from those systems versus the gains from ERP systems. Multiple estimations from software vendors cannot be simply accepted but must be blended with experience in implementation. Many software vendors quote a 1:1 ratio between software costs and consulting costs. However, this is not the extent of implementation costs. Software vendors only consider their consulting costs (although sometimes they add in training) and do not include internal resource costs for the implementing company. Generally, consultants from the software vendor provide the best consulting value. As soon as an outside consulting company is involved, the costs of the implementation go up. Therefore, to account for all of this, Brightwork Research & Analysis uses higher multiples if a client states that they are using a major consulting company, and companies performing TCO analysis themselves should do the same. Most vendors would not be happy with the implementation duration estimates developed by Brightwork Research & Analysis. However, these estimates are based upon years of analyzing how applications are actually implemented, which is far from the optimum values that are often quoted. We have performed the research, and the statistics are clearly on our side—enterprise software implementations take much longer than is generally assumed, not only by the software vendor but also by the project management of the implementing company. If sixty percent of ERP implementations fail, and if the vast majority of ERP implementations miss their deadlines by significant durations, why are TCO estimates still based upon assumptions that do not include these very critical factors? Ninety-six percent of ERP implementations include moderate to extensive customizations.

The Importance of TCO Customization

Customization results in high implementation costs, high continuous improvement costs, and high maintenance costs. Other software categories have various degrees of customization— almost always less than ERP—so ERP should receive the highest bump for coding-related implementation costs. All TCO estimations are based upon some type of duration, that the company uses the application. While different software categories have different average usable durations, there is really no perfect way to estimate this value, and it is difficult to know how long the application will be in use in the company.

Furthermore, an application that does not work very well can be kept too long—often for political reasons—while an application that is working well can be replaced due to issues that are related to what happens to be popular at the time. Support resources include everyone required to support the application: technical, functional, and management. It should never be assumed that the support “load” on internal resources is equivalent, even between applications in the same software category. There is a marked difference between vendors—and the degree to which the applications have been designed to be maintainable. This maintainability can be everything from how easy the application is to use (its usability: more usable applications require less hand-holding to accomplish tasks) to how straightforward it is to update its master data.

Specific Versus General TCO Calculation

Developing TCO estimations is the difficult part. The more interesting part is actually using the TCO, as there many varied uses. TCO can be used specifically or generally. For instance, once one has a handle on TCO for an application area, the TCO can be used to make future decisions after the purchase has been made. After TCO is developed, it can be put to use in supporting decision-making in a variety of ways. Companies should really have TCO analyses performed for all of their applications. It is quite common for companies to make decisions to extend the use of their ERP system in some area, usually functionality that is known to be mediocre; however, the decision is driven by the desire to “get more value from our ERP system” or to “leverage our ERP investment.” These can seem like desirable goals until one begins to look through the lens of TCO. Overall, “leveraging” any current software, be it ERP or other software that the company owns, will not typically save the company more than fifteen percent of the TCO of the functionality in that area.

Strangely, companies do not estimate their probability of success prior to deciding which project to fund. Instead, they take the naive assumption that all projects will succeed, even though IT projects have a high failure rate, even if the exact failure rate and the definition of failure is most often not specified.

TCO Versus ROI

It is important not to confuse TCO with its close financial cousin: return on investment (ROI). Below is a definition of how TCO and ROI are related to one another.

TCO analysis is not a complete cost benefit analysis, however. TCO pays no attention to many kinds of business benefits that result from acquisitions, projects, or initiatives, such as increased sales revenues, faster information access, improved operational capability, improved competitiveness, or improved product quality. When TCO is the primary focus in decision support, it is assumed that such benefits are more or less the same for all decision options, and that management choices differ only in cost. Encyclopedia of Business Terms and Methods

In actuality, few software products provide identical benefits, and the more complex the product, the more variability between each alternative return on investment (ROI). The simplest items commodities are the easiest to compare, precisely because their properties (in this case physical properties) are uniform. This is why silver, tin, grain, cotton, etc. can be sold on commodity exchanges. In contrast, enterprise software is complex; each application has different implications for how well the software meets the business requirements in terms of functionality, effectiveness of the user interface, ease of integration with other applications, amount of support needed, and hardware requirements and those are just the major items. Frequently the variability or score of these different items is unknown during the software selection process.

How ROI and TCO are connected.

The Connection Between ROI and TCO

TCO is the base value for ROI. A TCO must be calculated before an ROI can be calculated. ROI is the formalized analysis of the universal ratio between costs and benefits, which I referred to previously, and is focused on both the revenue and the costs of an investment. Unlike TCO, ROI takes into account the benefits from an acquisition. This is brought up in a quote from Ian Campbell of Nucleus Research:

“The real problem with TCO is that it’s a metric that can’t be used to make a buying decision. TCO assesses costs without regard for the benefits. We buy based on value, and I would challenge you to think of an item you purchase in your daily lives based solely on lowest costs.”

While this is true, the ROI is based partially upon the TCO. Therefore, while I don’t propose that TCO is the end point of the analysis, it is important to know the TCO before moving on to other analyses. Secondly, the ROI of enterprise software is quite difficult because it means estimating the financial returns from software, which is an extremely complex endeavor. Calculation of ROI is complicated by the fact that the specific areas of functionality that will be leveraged by the company must be estimated, and then the net benefit of that functionality must be projected. This is the only way that we know of providing a differential ROI between competing applications in a software selection.

The General Understanding of ROI

It is interesting to read articles about ROI, and after reading through them just to find essentially nothing explained of how to estimate ROI. There is actually far less published on ROI than TCO. In one way this can be seen as curious as so much is spent on enterprise software, but it is considerably less curious when one appreciates how difficult it is to calculate an ROI. And, in fact, once you dig into the detail of ROI estimations, it turns out that there is not much there. Furthermore, there are both explicit and implicit benefits to software, and the ROI estimations that we have evaluated will bring up the explicit or the hard benefits only, as it is often considered too difficult to ascribe a number to the implicit or soft benefits.

Therefore, instead of producing ROI estimates, Brightwork Research & Analysis applies a rating system to the application, which is used in conjunction with the TCO estimate. These are the following:

1. A Functionality Score
2. An Implement-ability Score
3. A Usability Score
4. A Maintainability Score

This results in a composite score, which the client can weigh toward the factors that are most important to them, or they can simply apply an even twenty-five percent weight to each score. This still allows the value of the software to be estimated, but without the necessity for an ROI calculation. The combination of the TCO and the application’s composite score captures the multiple dimensions of the application and is an effective approach for making a determination on software.

A company typically looks for the highest ROI, and simply the lowest TCO. For example, an application with a higher TCO than a competing application can also have a higher ROI, making the product with the higher TCO the preferred choice.

Net Present Value and TCO

Financial analysts, and anyone who has taken a finance course, are familiar with performing net present value calculations (NPV ). NPV discounts the future benefits and future costs into present day numbers through the introduction of some interest rate, often the company’s cost of capital. We do not add net present value calculations along with TCO estimations. If we did, it would shift the emphasis somewhat to applications that have lower up-front costs, such as SaaS solutions that charge for software on the basis of a subscription and have low implementation costs. However, the TCO advantage of SaaS solutions based on TCO in comparison to on-premises applications is already overwhelming and would simply show further differences.

How NPV Can be Calculated

NPV can be calculated once the discount rate the company wants to use is applied, but in most cases it will not make a difference between competing applications within the same software category, or software categories that are competing with one another for funding. And actually, it’s rare to include NPV in the TCO estimation. NPV is the most accurate approach, but TCO estimations tend to contain so many assumptions that NPV tends to further complicate the analysis (as well as the reader’s ability to understand the analysis). Secondly, as I will discuss in more detail, when the TCO methodologies are normalized (that is, adjusted to be comparable), then TCO from different sources can be used quite effectively; this improves the overall usability of TCO for the broadest decision-making. But, if other sources of TCO do not use NPV, it does not make a lot of sense to use NPV for one’s own TCO estimation.

Thirdly, TCO is essentially a comparative exercise. When TCO is calculated and a company bases a software, hardware, or services purchase on that TCO analysis, they do not actually have to cut a check for the TCO amount when they make the purchase. The TCO estimation is a guideline for how the future costs should work out. Therefore, as long as the various TCO estimations have a similar basis, or are normalized, precise adjustments such as NPV are not necessary.

The Purchase Price

The purchase price is referred to as an explicit cost, meaning the cost is published, obvious, and easy to include when making a decision. However, a purchase implies many other costs, which may or may not be known at the time of purchase. Total cost of ownership attempts to combine all of the different categories of costs into one number and use that number for making decisions. TCO can change a decision about a purchase that, at fi rst glance, may seem less expensive than other alternatives but in the end is actually more expensive.

TCO is discussed in the abstract but is rarely calculated in reality. If you think back upon all of the purchases you made throughout your life, how many of them included the TCO on the price tag, along with the actual purchase price? There is a good reason (or good reasons) for this and they go by the names of sales and marketing.

Sales and Marketing and TCO

The primary objective of sales and marketing is to increase sales. Sales decrease when customers come to understand the TCO of their purchases. Every purchase has a particular ratio: the ratio of the costs to its benefits. In the customer’s mind, the lower this ratio (as this ratio is perceptual), the more likely it is that they will make a purchase. Sales and marketing work their magic by increasing the perceived benefit of a purchase while decreasing its perceived cost. Having the prospective customer place the purchase on a payment plan can further reduce the sting of the most explicit cost: the initial purchase price.

The last thing a company wants is for their prospective customers to know the total cost of an item. The one exception to this rule is if the vendor has a study that shows their product or service has a lower TCO than that of a competitor’s product or service. Some software vendors create a white paper that shows the results of a TCO analysis within their software category, and lo and behold every occasion they, of course, have the lowest TCO! Even so, such advertisements of TCO in any context are rare.

What About Cloud or SaaS TCO?

SaaS software costs are easier to estimate than on premises cost because the purchasing company pays a monthly fee to the SaaS vendor. Some SaaS vendors throw in a set amount of free consulting – so the monthly fee covers a number of costs, which must be estimated for an on premises vendor. But it also makes comparisons between on premises vendors and SaaS vendors tricky. This is explained in the following quotation.

“Though TCO models are customary, the research found that the process lacks rigor and can be misleading. The study states: “As eCommerce and technology leaders factor in costs to their TCO model, they leave many stones unturned, which leads to a situation – intentionally or not – where the model can favor on-premise solutions versus those from the cloud. …Perceptions that cloud-based revenue share models are more costly than owning and operating ecommerce technology in-house are often misguided.” – Demandware

Demandware’s point is very straightforward. SaaS vendors charge a monthly rate – therefore their costs are more explicit. SaaS vendors charge for consulting as well, but because SaaS vendors already provide the server and the software ready for the customer to use, SaaS vendors have much shorter implementation times than do on premises vendors. However, if a TCO is not performed, it would tend to disadvantage the SaaS vendors and make the on premises vendors look like a better value than they actually are.

How to Make SaaS Fit with On-Premises TCO

To make the SaaS TCO estimation fit with the on-premises TCO estimations, it is necessary to add cost categories, which allows both TCO and on-premises solutions to be compared side by side, even though they do not have the same cost components. Simply comparing total costs for each delivery method is the most important issue. Effectively comparing on-premises solutions to SaaS solutions will be an important goal in the future as SaaS increases in popularity. In fact, our view is that a primary reason why SaaS is not more popular is that companies simply are not aware of the substantial cost benefits to SaaS-delivered solutions. There are significant differences in how long it takes to implement various applications. Some applications are naturally easier to set up. Other applications are simply designed to enable the vendor to say they have certain functionality. Instead of relying upon good software design, vendors rely upon aggressive sales, paying off Gartner for a good rating and/or their partnerships with consulting companies to get their software sold. Therefore, any TCO evaluation means spending time with the application, either implementing the application or using the application in a real-world setting such as testing the introduction of new master data into the application and rating the difficulty, or asking the vendor to demonstrate specific functions, which can prove out the implement-ability and usability of the software. One of the most commonly underestimated areas of TCO is the internal maintenance costs. The same issues that apply to the training apply here. Software that is difficult to implement is also difficult to maintain and vice versa.

The Dominance of Maintenance Costs in TCO

Companies that breathe a sigh of relief after a difficult implementation probably shouldn’t because the implementation difficulty is often—although not always—a good indicator of how much work will be required to maintain the application. Adjusting for the number of users is, of course, extremely significant for any TCO analysis, because the number of users is one of the most important drivers of cost. Applications scale in cost as the number of users increases. Therefore, a TCO that assumes 100 users will be higher with only 10 users—although the costs will not scale in a linear fashion.

In fact, other things being held equal, the higher the number of users, the lower the user cost per user. However, the per-user cost is not helpful for making value comparisons between software categories, because different applications vary quite significantly based upon the type of application. For instance, the categories of software that are most broadly used are ERP systems and reporting/business intelligence systems, while the categories of software that are used most narrowly include specialty applications such as supply chain planning.

Conclusion

This highlights a main theme of the book that while one TCO study can be useful—which is how TCO is primarily used—its usefulness increases greatly when TCO is performed for many applications and for many software categories. The most important conclusion from the research for this book is that companies do not do anywhere enough to leverage the power of TCO for their decision-making.