Electronic distribution: the rollout begins.

When we compiled our first ESD survey a year ago (Softletter, 10/10/96), electronic distribution was a fascinating innovation--the perfect channel for digital products, a way to bypass dysfunctional retail channels and "level the playing field" for small competitors. In the very near future, developers predicted, billions of dollars of software would flow through brand new ESD channels.

A year later, the revolution has moved to a new phase: old-fashioned trench warfare. ESD still promises a radical transformation of software distribution. But the implementation of ESD turns out to be complex and expensive. Building a good Web store is far from trivial, the tools and technology still have rough edges, and--surprise--giving away millions of trial copies doesn't automatically turn a company into the next Netscape or McAfee.

Still, all these hard lessons seem to have moved the market forward in a measurable way. Increasingly, ESD is not just wishful thinking--it's an emerging distribution channel that generates real sales (if not profits) for developers and publishers in virtually every segment of the software market. And a few of these pioneers have figured out how to make their virtual cash registers consistently ring up big-ticket sales.

With these trends in mind, we chose to focus our second annual Softletter ESD Survey on the practical issues of rolling out an electronic channels program. Through a combination of print and online polling (see page 2 for the demographics of our survey sample), we collected data on several key questions:

* What are the most popular configurations for ESD trial versions?

* How do ESD companies convert trials into sales?

* What's the relative contribution of a company's own Web store vs. online resellers?

* What's the prevailing solution to the online piracy issue?

* What's the likely impact of ESD on business and product strategies?

* And finally--when will ESD contribute at least a third of each company's total software revenues?

In talking with software developers, we've found these issues surface repeatedly in decisions about investment priorities and business models. Although many companies still seem to be collecting market research statistics to reassure themselves (or their investors) that ESD is more than a pet rock phenomenon, the more urgent challenge these days is to discover which approaches to electronic distribution really work, and which are the false trails. It's far too early to look for textbook answers to these questions, of course, but there are a growing number of pioneers who have found some relevant early solutions. This survey summarizes the experience of several hundred of these pioneers, and we suspect their conclusions will shed light on implementation challenges that other developers face.

TRIALWARE TACTICS: THE EMERGENCE OF TRY-BEFORE-YOU-BUY

Traditionally, software buyers have had to take a publisher's claims largely on faith: Retail boxes, direct mail promotions, and telesales conversations provide no hands-on experience of a product's features or performance. Worse, buyers tend to believe software is relatively expensive compared to other consumer and business products, so there's often a greater sense of risk-taking in purchase decisions.

So it's hardly surprising that the Web has become a giant try-before-you-buy channel. In fact, 69% of our respondents already offer a downloadable trial or demo version of their best-selling title, and another 16% have a version "under development." Only a modest 15% don't offer some kind of electronic trial experience.

But there is much less agreement about the best way to package a trial version. Currently, the most popular approach to trialware is to give customers a version of the product that "times out" after a period of usage. Another popular configuration is the so-called "crippleware" approach, which provides users with incomplete features or other limitations. In addition, many developers still rely on classic shareware marketing, simulations, moneyback guarantees, remote demos, and free downloads of obsolete versions.

[CHART OMITTED]

*The time-out approach: If a single trialware standard emerges, it's likely to be a full-featured configuration that expires after a user has spent time learning and testing the product. More of our survey respondents--122 companies--use this method than any other, and anecdotal reports from software marketers suggest that the conversion rate from time-out versions is usually higher than from other trialware methods.

Our data also points to a clear preference among developers for trialware versions that expire after a period of time rather than after a certain number of uses. Only 14 respondents say they limit the number of uses for a trial version; of these, six limit uses in conjunction with a time-out period.

It also appears that a 30-day trial period is the prevailing standard for most time-out trials. Products with longer expiration periods tend to be complex, high-priced titles that naturally require a longer evaluation cycle:

20 days or less (8 respondents)    8%
30 days (69)                       68%
45-60 days (13)                    13%
90+ days (11)                      11%

* The limited-feature approach: A fairly popular alternative to time-out configurations, limited-feature versions are used by 98 of our respondents. Some of these versions won't save or print; others create special file formats or add a message (such as "trial version") to documents. Among game companies, it's common to find trial versions that allow the user to explore only a few levels of play.

* The shareware approach: Classic shareware marketing (which many shareware developers have abandoned in favor of time-out trials) lets users download a full-featured, unrestricted version for evaluation. Customers who continue to use a product are asked to make a voluntary registration payment, and in return may gain access to tech support and other benefits. Shareware registration rates are notoriously low, but many advertising-supported Web sites now distribute literally millions of shareware downloads as a way to attract traffic--so it remains a reasonably popular ESD model, with 37 of respondents using some variation on shareware as their primary trial method.

* Other trial approaches: Seventeen respondents let users download a "non-functional simulation" of the product. Ten provide a full-featured version for an up-front payment, supported by a moneyback guarantee. Eight respondents let users download a non-current release, often with a built-in upgrade offer. And three rely on remote, Web-based demos.

CONVERSION TACTICS: HOW TO CLOSE THE SALE

Putting a trialware copy in the hands of sales prospects is only the beginning of the sales process. (Some of our respondents believe otherwise, including one who insists, "If the user likes it, he will buy it.") In fact, ESD is now widely regarded as an especially productive lead generation technique. ESD marketers use a variety of approaches to convert these leads to sales:

[CHART OMITTED]

* Upgrade offers: The most popular sales conversion technique is to incorporate an upgrade offer directly into the trial version. This approach, which 86 respondents say they currently use, doesn't require the developer to collect registration information during the download process. Like any traditional upgrade offer, however, the conversion rate depends on how compelling the additional features or services are to the user. If users feel the trial version is adequate, they probably won't spend money on the upgrade.

* Direct response methods: A large number of developers--79 survey respondents--now require users to provide basic registration information in return for "free" downloadable software. This contact information is sometimes used to generate direct mail (11 respondents), but usually the follow-up is a more immediate telesales call or other personal contact (68 respondents).

* E-mail campaigns: Trial users are often unwilling to supply much personal information, but most seem willing to provide at least an e-mail address. Forty-eight of our respondents use this e-mail address to make follow-on offers for upgrades, premiums, and special price deals.

* Voluntary payment: A substantial number of respondents--another 48 developers--say their only conversion method is to ask for voluntary payments by users. (Most of these respondents are traditional shareware developers.)

* Unlocking calls: Occasionally, ESD marketers require users to call for a special unlocking code to decrypt a trialware version, which may provide a chance to qualify the user as a sales prospect.

WEB STORES: WHERE ARE THE SALES?

Most online stores sell a mix of downloadable and physical product (CD-ROMs and full retail boxes). Although purists may not count physical product sales as ESD transactions, the reality is that both formats are an important source of revenue for Internet-based virtual stores.

Moreover, there are now two major business models for such stores: Some are direct response sites that publishers have built for their own titles, while others are true resellers that carry products from many different publishers.

In many ways, Web stores are the crucial link between ESD developers and publishers. Like real-world stores and catalogs, they have to attract customers and provide a selling environment for products, and then provide an infrastructure for processing orders, collecting payments, and delivering products. A successful Web store has to do all of these things reasonably well, and the glitches are conspicuous.

It's also not obvious how to measure the success of a Web store. What are reasonable benchmarks for traffic, sales, responsiveness, profitability, and the like? One fairly basic approach is to compare sales for downloadable and physical versions, and to see how the sales contribution of online resellers compares to that of publisher-owned Web stores. This may not provide a real measure of performance, but it does suggest where to expect the greatest sales contribution.

To develop these ratios, we asked our respondents to break out their unit sales (not trialware downloads) for the past three-month period from their own Web stores and from online resellers, and further to identify the number of downloadable copies and physical copies from each.

When we look at median numbers of units, it turns out that the typical publisher's Web store sells 200 downloadable copies every three months, compared to 120 sales of physical product. The specific sales numbers vary, of course, depending on company size, product category, and the sophistication of the Web store itself. But the ratio is the important number: It suggests that developers of Web stores should focus on delivery of downloadable versions, while not neglecting the contribution of physical formats.

We see almost exactly the same ratios among the online resellers: Downloadable versions typically contribute 100 sales every three months, compared to 50 units of physical product. Moreover, reseller sites are beginning to contribute a significant share of total sales volume--a dramatic change from our survey data just a year ago.

THE PIRACY ISSUE: COPY-PROTECTION ALTERNATIVES

The Internet has a well-deserved reputation as a haven for software pirates, and "unprotected" software seems especially vulnerable to copying and re-distribution. A group of ESD technology vendors have addressed this issue by creating tools for encrypting and locking downloadable products. With the help of these "digital wrappers," developers should be able to minimize the risk of piracy in online channels.

At least that's the theory. When we look at how publishers actually implement piracy protection, however, it appears that most companies rely on the same methods that prevail in other channels--in essence, trust and serial numbers:

* Voluntary compliance: By far the largest number of respondents, 250 companies, say they use "no protection" on their downloadable versions and rely on "voluntary compliance by customers."

* Serial numbers: A large number of respondents, 59 companies, say they "track serial numbers of downloaded version" as a way to identify pirates.

* "Digital wrappers": Fifty-six respondents say they rely on encryption and unlocking technology to protect their ESD titles. However, the leading technology vendors in this area are clearly not familiar brand names: Few respondents could supply the name of the wrapper company they used, and several mentioned compression tools (like PKZIP) that don't actually provide copy-protection.

THE LONG-TERM PAYOFF

What are the compelling benefits of ESD? We thought it might be useful to see how developers and publishers view this question, which in turn has an obvious impact on implementation and investment decisions. So we gave our respondents a checklist of ways that ESD might "change their business and product strategy."

A remarkably large number, 150 respondents, picked "more direct interaction with customers" as a major strategic benefit for ESD--well ahead of such benefits as "easier to compete with market leaders" competition (109 respondents) and "more revenue from add-ons, data, services, etc." (78 responses). In a market that relies more than ever on installed-base customers, the Internet has tremendous potential to improve communication--and it's reassuring to see that so many publishers already see this potential.

Our respondents also pointed to a good many other benefits that we hadn't included in our checklist. Various types of cost savings--from COGS, marketing, and support--were mentioned especially often, along with faster release cycles and access to international markets.

THE END OF THE RAINBOW 

When will electronic distribution become a major factor in software marketing? For many developers, the future is already here: 15% of our sample base say their primary sales channel is the Web (see page 2). Similarly, 17% report that ESD already generates more than a third of their revenues. Moreover, some of these ESD pioneers are substantial companies--not just garage-shop entrepreneurs.

Still, the majority of software companies haven't yet seen a cascade of dollars from their ESD investments. For these companies (and for newcomers who are still exploring electronic channels), the issue of timing is critical. It's pointless to build a sales channel before most customers have accepted ESD; it's even riskier to miss the first wave.

To help answer this question, we asked our survey respondents when they expected ESD to produce at least 33% of their company's revenues. The answer suggests that hockey-stick growth for ESD isn't far off. For the sample as a whole, the median prediction is 1999. Smaller companies tend to pick an earlier date, while larger firms have more modest expectations. (The median for companies with under $1 million in sales is 1998; for those over $10 million, the median is the end of the year 2000.)

We asked exactly the same question in last year's ESD survey, so it's possible to compare the two sets of predictions. One conspicuous difference is that expectations for ESD seem to have slipped by a year (these are projections from software companies, after all): In 1996, the median year for reaching 33% of sales was 1998. We suspect the schedule slippage is simply further evidence of the industry's experience with electronic channels: Radical change is still on the horizon, but getting there will take a little longer than most developers expected.

THE DEMOGRAPHICS OF ELECTRONIC DISTRIBUTION

The 1997 Softletter ESD Survey is based on data supplied by 332 software developers and publishers, who primarily serve personal computer market segments. (As a point of comparison, last year's ESD survey attracted 310 respondents, with roughly similar demographics.) This sample is characteristic of the overall makeup of the software industry:

COMPANY SIZE: Although most of the industry's largest PC software publishers supplied survey data, more than three-fourths of our respondents report annual sales of $10 million or less:

Under $1 million (138 respondents)    45%
$1-$5 million (71)                    23%
$5-$10 million (28)                   9%
$10+ million (69)                     23%

PRICE CATEGORY: Increasingly, pricing levels have become an important indicator of market segmentation. Our questionnaire asked for the list price of each company's "most popular title," and the answers suggest that ESD companies currently sell products across a wide range of price points, especially in low-end consumer segments:

Under $50 (109 respondents)    36%
$51-$150 (65)                  21%
$151-$600 (66)                 21%
$600+ (67)                     22%

MARKET SEGMENT: We asked respondents to identify the market segment that "contributes the largest percentage of your sales":

Consumer/home office (115 respondents)    36%
Small & mid-sized business (81)           25%
Large corporations/Fortune 1000 (76)      24%
Industry-specific/vertical market (47)    15%

SALES CHANNEL: Since ESD is itself a distribution channel, a company's traditional approach to channels is bound to have an impact on how it implements electronic sales. We asked respondents to identify their "primary sales channel," and we also calculated the median list price for products in each segment. It's significant that 15% of our respondents already use ESD as their primary sales channel; it's also an interesting sign that 14% rely on shareware distribution:

Direct sales force/telesales (102 respondents)    31%    $495
Reseller/VAR/third-party catalog (96)             29%    $99
Electronic/Web site (50)                          15%    $59
Shareware (45)                                    14%    $30
Direct mail/direct response advertising (34)      10%    $199

Note: Not all respondents provided data for all survey questions. Percentages may not add up to 100% due to rounding.