How to keep support costs under control.
"If you figure the average support call is about $25 [and] if prices stay in the $100 to $200 range, you'll have to unbundle support. That's absolutely the case." --Philippe Kahn, Windows Summit '93
At the Windows Summit earlier this month, Borland chairman Philippe Kahn warned that the current price wars have "absolutely" doomed free technical support. No company, he says, can afford to give away support for products that sell at the deeply discounted prices that now prevail in the marketplace: Soon, support will have to be "unbundled" and sold as a separate service.
Well, we disagree. To be sure, almost every major developer has seen the cost of support virtually explode over the past year or two. The old rule of thumb was that tech support represented about 6% of revenues; today, the percentage often exceeds 10%--even for companies that are still selling full-price products.
But this cost explosion isn't necessarily bad news: The rising demand for support is a symptom of substantial numbers of genuinely new users, who have been lured into the market through a combination of aggressive discounts and glitzy new products. Low-priced software isn't inherently expensive to support (as Borland proved with its original Quattro Pro competitive upgrade, which appealed primarily to experienced spreadsheet users). Instead, support costs escalate when there are lots of neophyte users, new technologies, new applications--the same forces that have always fueled the industry's continuing growth.
And if the goal is to keep attracting these new users (rather than selling the same applications to the same installed base), then charging for support is precisely the wrong answer. Free support almost certainly reduces the perceived risk for users who are thinking about trying out new technologies or even buying an upgrade. A lot of paid-support advocates conveniently forget that new software has a reputation--true or not--for being confusing, buggy, or otherwise flawed. Eliminate the free-support safety net and a lot of risk-averse users will probably decide to stay out of the market.
There's also a practical problem with paid support: It almost never generates much money. We've collected a good deal of statistical data about tech support call patterns (Soft-letter, 2/18/90), and it's clear that the majority of calls are inspired by difficulties with newly-purchased products. (On average, 66% of calls occur within three months of purchase, 83% within six months.) It's hard to imagine eliminating free support during some reasonable warranty period, so the actual revenue potential for fee-paid support is probably no more than 15%-20% of calls.
In fact, paid support plans rarely reduce the overall cost of operating a support department, and sometimes increase those costs. Typically, paid support implies a premium level of service, which costs significantly more to deliver and market than ordinary support services. (The real payoff from such plans is that they appeal to corporate users, who want at least the opportunity to buy direct access to a developer's best technicians. But that's another subject entirely.)
In short, "unbundling" support isn't going to solve the cost crisis. At the same time, however, support doesn't have to be a bottomless pit: Almost every major software company--including Borland, incidentally-- has been experimenting with methods for improving support productivity, reducing call volumes, and otherwise cutting costs. So far, there are at least a half-dozen strategies that promise to yield important payoffs:
* Hire more machines, not more people: Tech support automation--in particular, fax-back and interactive voice systems--is a hot issue these days, and with good reason. Typically, such systems offload anywhere from 5% to 15% of total call volume, and (like bank ATM machines) they're generally perceived as a service enhancement by end users. Based on present usage trends, our guess is that within the next five years, at least a third of all support calls, upgrade orders, and other routine inquiries will be handled as automated transactions.
* React faster to usability problems: When tech support reports a major bug, most software companies immediately send in the R&D SWAT team. Yet the really costly support problems these days tend to be usability issues--installation procedures that are hard to follow, features that don't work properly, documentation that baffles even expert users. Letting these problems linger on is expensive, but tech support departments rarely have the authority to make the necessary changes. Microsoft may have figured out the best strategy: The support organization tracks calls by product, and then charges these costs back to the P&LS of individual product groups.
* Don't waste money on irrelevant services: Support organizations often spend huge chunks of money on services that customers barely notice, then scrimp in areas that are more critical. (For instance, corporate users often admit that toll-free telephone access is a very low priority.) There's no magic to figuring out what customers think is important; surveys, focus groups, and face-to-face meetings can provide plenty of good advice about how to set service priorities.
* Give technicians better information tools: When we surveyed tech support productivity levels last year (Soft-letter, 7/7/92), we were surprised to find that technicians typically handle only about 20 to 25 calls per day. One major factor is the amount of relatively unproductive time that technicians spend learning new products and researching obscure problems. Here, there's a growing consensus that knowledgebase systems, by reducing the need for in-depth training on every product and eliminating redundant research, can have a dramatic impact on productivity. If end users are allowed to tap into the support knowledgebase, moreover, there are usually secondary cost savings as well.
* Keep track of customer names: Rather than deny support to "unregistered" users--that is, software pirates--many developers now use free tech support to capture names and demographics of virtually any user who calls for support. The goal: To build a rich database of potential buyers for upgrades and aftermarket products. Maintaining this data is expensive, but every analysis we've seen suggests that future "life cycle" revenue more than offsets the costs of providing open-ended support and capturing detailed customer demographics.
* Outsource "sunset" and upgrade support: A few large publishers have recently (and very quietly) turned over some of their support calls to third-party service companies, such as Corporate Software and Software Solutions Inc. Most often, the outsourced products are "sunset" titles that the company no longer expects to keep marketing, or upgrades that would otherwise overwhelm the in-house support department. On a percall basis, outsourcing rarely seems to generate direct savings, but it does eliminate some of the inefficiencies of rapid and uneven growth.
|Printer friendly Cite/link Email Feedback|
|Date:||Mar 31, 1993|
|Previous Article:||The 1993 Soft Letter 100.|
|Next Article:||InfoCorp: a new user profile?|