Telegraph Hill , CA
September 14, 2005

Whither Salesforce.Com?

What value does Salesforce.com deliver? Hint: It isn't just an outsourced data center.

A Salesforce Customer

Who are the customers of Salesforce.com? Well, poor little B2B Analysts, Inc., was one. A new partner urged us to coordinate sales efforts; $995 was within range for an experiment with this; and as soon as the credit card charge hit, we were in business.

We found that we had joined a lot of other companies we knew. A competitor. The people down the hall. Many of the small software companies we cover.

In this piece, the fourth in a series, I want to talk first about the value we and other customers got. In this series, I am arguing that best-in-breed application companies can only survive if they learn how to get into the value delivery business and out of the software delivery business, the business that the large application companies are in.

Salesforce.com is a showcase for this argument (though I'm sure several readers will wonder at the application of the word "best" to the rather thin suite of products that the company provides). As we shall see, it delivers a unique and interesting value that the market has responded to with overwhelming eagerness. And it bids fair not only to survive, but to grow by delivering more of the same.

I'm going to spend a good part of this piece on Salesforce.com's SFA product, which I will call "Salesforce" (no com). You'll need to understand how the product is put together and why if you're going to understand the value delivered. I'm then going to talk about the value, which is not what most people say it is.

I'll conclude with a discussion of where Salesforce.com is going. I've just returned from Dreamforce, their user conference; it turns out that they are going to grow, surprise, by becoming a platform company.

Training Wheels for CRM

Salesforce.com sells hosted sales force automation software on a subscription basis.

There is nothing original in this. Salesforce.com didn't invent On-Demand, IBM did (and Larry had a pretty deep insight into the idea a decade ago). They didn't invent hosted software or hosted CRM; hosting emerged in any number of places. They didn't even invent the idea of selling software by subscription.

Yet somehow, to the intense irritation of many software CEOs and industry observers, their particular version of hosted CRM has found a large and almost entirely unsuspected market.

Just what the market likes so much was originally a bit of a puzzler to an old software hand like myself. The product itself is a rudimentary account-contact-opportunity system, prettily wrapped and accessible over the web. The software wouldn't ever win any functionality wars, and even within the account-contact-opportunity structure, I personally found it difficult to work with.

In fact, our overall experience was instructive. We tried to use it. But we struggled. None of us were quite used to being disciplined about entering data, and having to have an Internet connection to enter the data made things worse, not better. (It always seemed that it was on planes that we wanted to enter in business cards.) I don't know if it was us or the interface, but our data was never accurate or consistent, and that meant that the reporting, which we found difficult to do anyway, had less and less meaning as time went on.

After a couple of months, we stopped using it. We'd get a cheery e-mail every quarter from Salesforce.com, with a report on HOW MUCH we used the system. Every number was a zero. Kind of amusing.

Amusing, that is, until the year came to an end and another $995 appeared on the credit card. "You signed a two-year contract," they said. "No we did not!" we replied. This went on for about two months. Then they give in and credited the $995.

I should probably never have ventured the $995 in the first place. As my friend Andy Zoldan says, SFA systems--Siebel, SAP, Act!, whatever--are "accounting systems for salespeople." If you approach SFA the way accountants approach GL, you can get very tight coordination and control of sales activities. If you don't, well, you get the experience I should have anticipated.

We've since made it a practice to compare our experience with those of the companies we work with. The reports have varied widely. A mid-size application vendor with close ties to Salesforce.com's management said, "Yes, it's great. It may only have 60% of Siebel's functionality, but that's all we really need anyway. We haven't had anywhere near the problems you remember we had back at XXX (very large application company with a failed Siebel implementation)."

A 30-person competitor, "We used it for a while. We dumped it." A 10-person manufacturing company, "Fabulous software. So much better than Gold Hill (a desktop contact management system)." A small application company. "Yes, we use it. We don't like it, but what are you going to do? Do you know of something else?"

A staple of Salesforce.com events is the happy user, the person from companies large and small who get up and tell us how they made the product work and work well. Even these accounts vary widely, but if there is one thing they seem to have in common it's that somebody in the organization insists on driving everybody else to put in the effort it takes. Comments like, "If it isn't in salesforce it doesn't exist," or "If you don't like the system, you probably don't need your job," abound in these presentations.

In companies like ours, where no one was convinced of the value of maintaining contacts the way you maintain the ledger, there was little or no value received. A majority of companies seem to lie in between, sometimes ending up with slapdash implementations and sometimes feeling their way toward approaching the problem with discipline.

Oddly enough, the successes can be of any size. I've talked to large companies that just wanted to get their leads and lead management under some kind of control and did really well with the product. And there were pure startups, two people, who grew very rapidly with Salesforce "and couldn't have done so without it."

Also surprisingly irrelevant are the product's limitations. In talking to people, there were many acknowledgements that the system didn't do "everything." But it was never clear that these lacks made the unhappy customers unhappier or prevented the happy customers from doing what they want.

What limitations? Well, consider that Salesforce.com is an "enterprise-level CRM" product where you can't do any of the following:

• Take an order from a customer, get a price for a customer, or even figure out what the customer already has ordered from you.
• Use Outlook to send an e-mail or schedule an appointment with a customer and have the e-mail or appointment show up seamlessly in Salesforce.
• Execute any kind of program or script. In Salesforce.com, you can enter data and you can get data out, that's it.
• Customize the system, except by defining new fields or tables or by populating drop-down lists.

If you've actually built enterprise applications, you are probably shocked by the last two items. Salesforce, it turns out, is little more than a system for doing innies and outies, grabbing information, storing it, and retrieving it: SQL with an interface. Far from harnessing the power of a computer, Salesforce has deliberately eschewed it.

If you're merely a user of these systems, you are probably not shocked. But you might sense that you are being asked to do work that might be done for you in more "advanced" systems.

For this user, the lack of scripting (programming) didn't matter much, but the lack of seamless integration with Outlook was a killer. If you wanted to send an e-mail with an attachment to a customer, well, you need to move over to Salesforce and create and send it there. If you wanted to schedule a meeting, ditto. There are contacts within Outlook and then contacts within Salesforce. I only want one home for my calendar, e-mails, and contacts. With Salesforce, I had two.

We tend to think of utility and functionality going hand in hand. The more of one, the more of the other. With SFA, where getting people to use it correctly is such a problem, you might think this is particularly true. As Pat House (of Siebel) once said to me, "We couldn't really succeed with this until we put enough stuff in to make the salesperson happy about using it." But Salesforce turns this idea on its head. Less seems to mean more.

One way of explaining this is to say that Salesforce with its CRM-light functionality, is a kind of training wheels for CRM. At a low cost of entry, customers could try out SFA, get some practice, see whether they liked it, and later graduate to "real" CRM, whose functionality made using it pay off more.

But after talking to a lot of customers, I've come to see that a CRM-light product may be what the market wants. Yes, there will be a fair number of customers who give up on the program and others who graduate to something bigger. But Benioff appears to have remembered that most enterprise-level application customers severely underutilize their systems. For such customers, Salesforce is as much as they need.

If you think about it, it's an oddity of the software market that the leaders are the ones who sell the luxury products and the "niche players" sell the product for the masses. It's our normal consumer experience for the mass-market product to establish itself with heavy spending on customer acquisition and to entrench itself not by spending a lot on improving the product, but by exploiting the brand.

As this week's Salesforce.com user conference showed, this last is indeed the direction Salesforce is taking.

Where Is the Value

But before we talk about the future, let's step back and look at the value Salesforce.com is uniquely delivering.

The following seem to me the key points where Salesforce is doing something quite different from the competition and thereby delivering a real value unique to them.

• They are delivering the retail buying experience to the purchase of enterprise software. The software is enterprise-level, that is multi-user and multi-location, but you can try it out free and buy it with a credit card.
• They are delivering the retail customer experience to the use of enterprise software. The buyer is the user, and the user can set it up any way the buyer wants, any time they want, without having to get IT involved.
• They are giving the user the emotional satisfaction of thumbing their nose at IT. More on this in a second.
• They are giving buyers at the intermediate level of a corporation a way of buying that doesn't impact capital budgets; they can use their signing authority to buy, without having to consult a committee.
• By limiting the potential of the product, they are also making the delivery of the available value far more reliable. Implementations are less complicated and are far more amenable to improvisation, er, continuous improvement.

There you have it. Modest? Perhaps. But big enough to appeal to a surprisingly large and varied audience. "It's simple. It does what we want. And we control it."

This value proposition has opened up a large and hitherto unsuspected market. The largest segment of this market numerically is composed of the leaders of small groups of salespeople, either the VP at a small company, or a pod manager at a large one. But large companies can "get it" too. There's a division of a $1 billion company in our area whose sales VP appears at local Salesforce.com meetings. He was tired of having customer information scattered all over the place and loved the attention that Salesforce.com can still afford to lavish on its few large customers. The company was too big to go pay with a credit card and hide the sale in its operations budget, but IT was happy enough to go along. They felt the value, too.

Even at companies committed to another platform, the ease of access to the product and the relative simplicity of implementation can lead them to do what we're taught is the unthinkable. Millipore, for instance, is a long-time Oracle reference client who put 500 users on Salesforce; there are equally long-standing and larger SAP customers who have done the same.

But you have to buy into the value proposition. A large Oracle shop like Cisco, which also committed to Salesforce was never likely to have an easy time of it. Cisco is notorious for buying small (or large) software packages, customizing the heck out of them, and then complaining, as one vice-president did to me, that the packages are pretty weak. Well, with Salesforce, this strategy is going to make for problems, because you don't get to customize the heck out of it. I would guess that these problems are at the heart of Cisco's issues right now.

One might think that with so modest a value proposition, Salesforce runs a severe risk of being commoditized. Perhaps. But that can't happen until the other On-Demand vendors have learned the lessons that Salesforce has to teach.

IBM and Siebel, for instance, seem to think that On-Demand is about availability; their aim has been to build a product as good as Siebel that you can log into over the Internet. (They've done this by the way.)

But the On-Demand value prop is not about availability. It's about control. Salesforce.com is all about taking control away from IT and giving it back to the business users.

You see the power of this value proposition most starkly at the Saleforce.com road shows, where selling against IT is the order of the day. But look around you. To business users, IT too often means delay, it means unresponsiveness, it means "No, no, and again, no." No wonder a sell against IT has so much power. For people who feel this way, On-Demand means they no longer have to depend on a service they don't like, don't trust, but heretofore had to use.

Unless they need to integrate, of course. Perhaps ironically, one of the first big improvements that Salesforce.com made to their product was a set of integration tools. It's turned out, though, that IT departments have been willing to swallow hard, go along, and learn.

It's probably been harder for the vendors to do the same thing. I went to many IBM and Siebel road shows, too, and never once did they seem to realize that On-Demand to the core Salesforce audience means selling against IT. No, they stood up, as potent symbols of IT as there are, and talked about features and functions.

Beyond SFA: Dreamforce 2005

At the Dreamforce conference, at least half the questions at the analyst lunch were about where Salesforce.com goes from here. Some questioners seemed to feel that SFA was a lucky break and hence unreplicable. Others expected Salesforce.com to begin going into the ERP market.

Most of the questions, though, didn't seem to appreciate the technical issues.

Remember, Salesforce is a push and fetch system only. And remember, too, that this design is essentially forced upon them by the fact that their application runs in a single instance on a single server (actually a set of servers). If that server had to be on call for any function that users could dream up, it would die frequently. But when Salesforce limits the users to using it as a data server, it will behave reasonably reliably.

Some smart reader may ask, "But can't you customize Salesforce?" Yes. But all of the customizations operate under the same restrictions. You can parameterize behavior a little bit (with drop down lists). You can create new fields and even new tables. With the new releases this year, you can write your own custom program that looks like Salesforce.com, is reached via Salesforce.com (that is, the user logs into Salesforce, selects a tab, and is transferred to the program), and does anything you can do with a compiler. But that "anything" doesn't run on Salseforce.com's server. It runs on yours.

So most of the questions missed the mark, because the questioners didn't understand the restrictions.

What does one do with this good-looking system and this brand? Well, an application vendor would say, "You do more than store and retrieve." Not Salesforce.com. In my obligatory five minutes with Marc Benioff, I commented on this, and his reply, which makes perfect sense, was that you store more information. "The more information we can process, the bigger we will grow," a succinct and accurate statement.

As I see it, Salesforce is inventing a new product category, which I, with my tin ear for marketing, would simply call "business information services." The category offers business users a place where they can safely store and retrieve structured information that they and their colleagues use in their business (like customer names), using an interface that is pleasantly simple.

A business information service is not a file system or a document management system, because it works with structured information, that is, information that is typically stored in discrete fields in a database system. And it's emphatically not a transaction (ERP) system, because it doesn't even try to do anything that could require that it be a system of record. (If I'm talking in code, I apologize.)

With a business information service, you might think that the biggest advantage is the fact that it's hosted, so anybody, anywhere can get at it, without having to worry about where in the file system it is. But I actually think that the big advantage is that it's structured using a manageable interface. There's an awful lot of information like this in organizations, which ought to be more accessible, but because it needs to be protected, it ends up behind the iron walls of the ERP system or buried on some spreadsheet.

At the recent Dreamforce meeting last week in San Francisco, Salesforce.com did not use the term "business information service," which probably helped its stock price. But the company made it very clear that this is in fact what it was doing.

The big announcement was something called AppExchange, a new section of Salesforce that allows users to test drive, review, and buy software that "runs" on the Salesforce.com platform (with the aforementioned caveat that the only actual processor time supplied by Salesforce.com is used storing and retrieving information kept on the Salesforce processor).

Rather than trying to find other business information that it can store using a structured platform--the obvious candidate might be HR information, but there are obvious problems with that, too--Salesforce.com has decided instead to let other people invent their own services, but force the user to "go through" Salesforce in order to get the information that the services provide.

In effect, by opening up the platform, Benioff and company can focus more on exploiting the idea of a broader information service, where users control the information and its mode of storage. Benioff and company are looking at models like Google and eBay, especially eBay, because they realize that they can't make this happen unless the user feels that they are joining a community. In a way, an apt analogy is the blogosphere. Blogging software gives control to the writer and not to the publisher. But that control would have little meaning if the blogger weren't joining a community, which later supports and maintains him or her.

Where can they go? Growth for Salesforce.com's version of On-Demand will be a three-step process. First, they establish that On-Demand is about getting out from underneath IT. Then, they establish that it's about getting a tighter connection between application provider and user (not only because the user has control, but also because the provider can continuously improve the product, testing usage on a daily basis and when the providerfinds a problem, pushing out a solution immediately). And finally, extending the range of the information service by establishing a community of users and providers.

This last needs to be done delicately. But I will defer a deeper discussion of the AppExchange to a later Short Take.

The Case Against Salesforce.com

Salesforce.com is riding a rocket, as one of its partners said. Old software hands have seen many such companies, most of whom, it turns out, got their rockets from NASA.

So what are the potential problems?

In discussions of this subject, two potential problems are always brought up. One, people argue that the market will commoditize quickly because the cost of entry was too low. Two, they argue that the customers would abandon their training wheels once they grow up.

The first never struck me as all that compelling. Salesforce.com is in the hardware rental business as well as the software business, a business that requires capital and technical oompetence. As for the product, the foregoing has shown that it is very carefully designed. That kind of care is not easy to replicate, at least not quickly.

The second has turned out to be, well, less likely than I originally thought. Exiting is never easy. (The fact that Salesforce spits out your data into a spreadsheet when you exit doesn't exactly encourage people to leave.) With ERP retention rates for outmoded products in the high 80s and low 90s, surely it would be hard for Salesforec.com's retention rate to fall below the 70s. In my first wave of irritation at my own experience, I thought it might be higher, but I was wrong.

Whatever the retention rate, it seems clear that Salesforce has the wherewithal to attract new customers faster than it's losing the disgruntled. There are 500,000 companies in the United States with revenues in the 9 figures, if I remember Econ 101 correctly. Many of them have people whose job it is to sell. Most still do not have any effective way to coordinate their salespeoples' efforts. And an awful lot of them like control and don't like IT.

So there is plenty of room.

But there are other problems. Consider the following:

Risk Tolerance. This is a company that has bet not only its existence but the assets of its customers on the utter reliability of a single computer installation, which it has decided to locate in San Francisco. In a few months, a mirroring site located on the East Coast will go on line. But the fact that management has been willing to gamble on there being no earthquakes for the last five years suggests that they are perfectly willing to run risks that might give others pause.

Amateur Hour. Operationally, everyone tells me that the company has improved over the time when that second $995 charge landed on the credit card bill. But it's still pretty clear that this company is still discovering the disciplines that are required of a company that wants to provide a retail buying and using experience.

Buyers Mature. As On-Demand becomes more accepted, buyers appear to be maturing and doing their homework. This may leave them less than satisfied with some of the limitations. Take, for instance, the claim frequently made that Salesforce is "integrated" with Outlook. Well, it is, in a sense (the details are on their web site). But in the last Salesforce meeting I went to, for the first time, potential customers showed that they had begun to understand that the integration they claim does not provide what people typically want, a single home for scheduling and e-mail.

Corporate Style. We all know that Salesforce.com is full of bright people who make decisions quickly and are willing to experiment. We all know that at least one of those people is not shy. This corporate style is refreshing and creative and it gets them far, fast. But we also know that not dissimilar styles dug holes so deep that companies like i2 and Ariba till can't dig themselves out, even years later.

Nowhere to Go. Finally, and perhaps most important, it may simply turn out that Benioff's insight into what really works in a mass-market SFA application is simply not replicable in other areas. You can't run a business information service if the only information that is suitably served up under theri model is information about contacts and opportunities.

What Salesforce.com Can Teach Us

Whether Salesforce.com succeeds or fails--and despite my $995, I wish them well--they have plenty to teach us about the business. Here are some lessons everybody can learn for free.

• End users see a significant value in having enterprise-level software that they can control themselves. I'm not talking about personalization; I'm talking about control of the setup and the use of the software.
• A best-in-breed company needs to remove the barriers that IT interposes between it and the end user or the buyer. You don't need to be a multi-tenant provider to do this.
• All best-in-breed companies need to think carefully about how they can deliver improvements without the need for IT to intervene.
• A recurring revenue business model is a good way of helping companies to justify the discipline and focus required to do the foregoing things on this list.

If AppExchange works, Salesforce.com may also have lessons to teach about establishing a community. But that is a subject for a forthcoming Short Take.

To see other recent Short Takes, click here for a listing.

Training wheels for CRM.