CIO

The Truth About On-Demand CRM

Despite the hype, the truth is that hosted solutions aren't going to take over the CRM world anytime soon.

Hosted, on-demand CRM is sometimes cheaper and easier to roll out than the software that lives on your own machines. But if you think on-demand means that all you have to do is flip a switch, you're dead wrong.

READER ROI

  • Why the rush to on-demand CRM
  • How to figure out whether on-demand is right for you
  • Why on-premise CRM remains the best choice for many

When Alex Marxer began looking at customer relationship management software, on-demand CRM wasn't even on his radar screen. As vice president of financial services for ResortCom International, a $US15 million business-process outsourcing company for vacation property developers and managers, Marxer was looking for an enterprise CRM system for sales, marketing, customer support, self-service and analytics. He needed software that was flexible enough to accommodate the changing needs of his sales and marketing staff yet would integrate well with the company's home-grown back-office applications containing all its customer contracts, invoices and financial transactions. So he had his sights set on traditional offerings from vendors like Siebel, Kana and Pivotal.

But then Marxer came across a hosted - or on-demand - CRM offering from RightNow Technologies that seemed to provide most of the functionality his business would need coupled with a particularly user-friendly interface. He was impressed with the price tag - just $US125 per user per month compared with the $US300,000 ResortCom would have to shell out for an on-site solution (not including implementation, infrastructure and support costs). So he signed a three-year contract with RightNow. "When we did the ROI calculations, it was an unbeatable value proposition," Marxer says.

Once the implementation began, however, Marxer ran into some problems. He wanted users to be able to launch RightNow applications as tabs within his back-office system. But that was impossible using RightNow's application programming interface (API) tools out of the box. RightNow sent a team to Marxer's San Diego office to work through a solution, which extended the implementation from the promised one month to three. Since ResortCom was entering its busy season (November through March), Marxer had to delay deployment until April. Since then, he has found that upgrading the RightNow software causes the integration with the back-office application to break. So he's had to hold off on upgrading to any new versions of the software - and forgo the valuable new functionality those upgrades would bring.

Marxer is one of thousands of executives who've made the decision to take a chance on the on-demand CRM model. It's a booming market - revenue from hosted CRM applications grew 105 percent last year, according to AMR Research. Small and midsize businesses and departments within larger companies have been drawn to these software-as-a-service solutions (payable on a monthly basis) because they're much cheaper than licensed on-premise software, which can cost anywhere from several hundred thousand dollars to several million up front. Salesforce.com, which created the model for hosted CRM in 1999 and developed a strong foothold in the mid-market, is now offering functionality beyond sales-force automation and trying to sell its product to much larger customers. And traditional CRM players like Siebel (now owned by Oracle) and RightNow have been forced by Salesforce.com's success to create hosted CRM solutions. Microsoft also recently announced plans to roll out an on-demand CRM product soon.

But despite the hype, the truth is that hosted solutions aren't going to take over the CRM world anytime soon. As Marxer found, implementing on-demand CRM software is not always as simple as vendors would have you believe. Customization can be problematic. Hosted CRM vendors' API tools cannot provide the degree of integration that is possible with onsite applications. Getting a hosted CRM system working shouldn't take as long as a traditional software package, but larger and more complex roll-outs can still take a year or more. And while the hosted option reduces the need for in-house technical support, upgrades can still sometimes be technically tricky, and ongoing business support of the software is critical. In addition, some companies with particularly sensitive customer data, such as those in financial services and health-care, do not want to relinquish control of their data to a hosted third party for security reasons. As a result, AMR Research predicts that by 2009, hosted CRM applications will account for only 12 percent of the total US CRM market.

"There's an expectation gap in the market," says Rob Bois, senior research analyst for AMR. Companies, he says, believe that "the on-demand model eliminates the up-front cost and effort required in implementing a CRM system; that it's just like turning on a switch. But the integration and customization requirements are not that dramatically different from traditional software when you get into more complex implementations. There's still a lot of work involved."

There are, of course, many situations in which the on-demand option will be the best choice: It should work well for companies that want to implement standard CRM processes, are able to use out-of-the-box data structures, have little or no IT support, or don't require complex or real-time integration with back-office systems. But as on-demand CRM extends further throughout the enterprise, it faces the same challenges as traditional CRM: ensuring widespread user adoption and integrating the system with other applications, says Rebecca Wetteman, vice president of Nucleus Research.

Some of the CIOs interviewed for this article chose an on-demand package, while others went with on-premise. They are candid about the pitfalls of either approach, but all agreed on one point: First you have to figure out exactly what your company wants from the CRM initiative and proceed from there. Price, they say, should be the last thing you consider.

"You have to have a strategic plan for what you want from your CRM initiative," agrees Paul Greenberg, president of The 56 Group and author of CRM at the Speed of Light. "Define your processes, figure out what your requirements are, decide who will execute on it. Then you can go through the costs of each model that actually meets your requirements and make a decision."

Page Break

When On-Demand Is Worth Considering

Hosted software is nothing new. In the 1990s, hundreds of ASPs sprang up to offer customers enterprise software hosted over the Internet. But when the Internet bubble burst, many ASPs went belly-up, leaving customers in the lurch. But Salesforce.com focused on the niche need for sales-force automation, refined its technology and began racking up sales among small and midsize businesses that needed the functionality they could offer but couldn't afford the multimillion-dollar price tag that accompanied full-fledged CRM implementations. And as the number of expensive failures in the traditional on-premise CRM space grew, so did interest in expanding the hosted model beyond simple sales-force automation - to a full-fledged system that could give enterprises a holistic view of their customers and allow them to better target their marketing, sales and customer-service efforts.

Traditional enterprise software vendors soon struck back, attacking on-demand on the grounds that it wasn't scalable to more than 1000 users. But as it turns out, size is not what really matters in determining whether a hosted CRM implementation will be a success - it's complexity. And complexity makes it harder to implement a viable CRM package, regardless of whether it's hosted or on-premise. "It isn't a scalability issue," says AMR's Bois. "Typically when you're talking about an organization that will have more than a thousand CRM users, you're talking about a broad implementation that will touch more areas of the company and will involve more business processes."

Companies seeking to adopt established, standard practices on a particular function like sales-force automation are more likely to benefit from a hosted solution, while those seeking to implement highly customized customer-management processes would more likely value a flexible on-site option.

Take SunGard Data Systems, for example. For Bettina Slusar, SunGard's senior VP of global accounts management, opting for an on-demand solution in 2002 was a relatively easy decision to make. Although she had a user base of more than 1000 to consider, her plans to drive standard processes in the global sales function of her $US3.5 billion data-centre company led her straight to Salesforce.com. Having grown through more than 100 acquisitions, SunGard's scattered and independently operating sales force hampered the company's ability to get an accurate and timely enterprise view of the sales pipeline. In addition, SunGard was looking only for certain aspects of CRM - sales-force automation and some marketing and campaign tracking. So Salesforce.com was a good fit.

But Slusar admits it's not for everyone. "If you want one big system that's going to connect all the dots together - from talking to the customer to signing the deals to connecting to the accounting system - [Salesforce] is not the answer," she says.

IT support was also a factor in Slusar's decision. SunGard's business is built on running data centres for other companies, but there is no centralized IT function for internal support. "We never really seriously considered an in-house solution for this. Our sales force spans the globe, and administratively it would have been a big headache to maintain the application and keep servers up and running with people coming online in Hong Kong and Chicago," Slusar says. In addition, she says, there was "no comparison in terms of pricing". Traditional software would have cost $US18,000 per user over the course of a two-year licence, and though Slusar will not reveal how much SunGard pays for the Salesforce.com systems each month, she says the cost is "magnitudes" lower, ranging between $US1560 to $US3000 per user over the same period (not including training and customization).

When On-Premise Is a Better Fit

At Qosina, a medical-components distributor, the basic business need was the same as SunGard's: creating an enterprise-wide snapshot of the sales pipeline. Yet this much smaller company - $US25 million in revenue - chose an on-premise offering in large part because of the complexity of its business processes, according to COO Gerry Quinn.

The company had been using an old flat-file, DOS-based contact manager called Telemagic for years and, in 2003, was ready to replace it with a more modern, robust CRM package. Qosina, which markets its products (such as tubing, clamps and valves) through trade shows, catalogues and Web sites, has an inordinately long sales cycle as the components it sells are eventually incorporated into products developed in the medical and cosmetics industry. To encourage the purchase of its products, Qosina sends out samples of its 5000 products (some costing less than a cent a piece) at an average rate of 300 to 500 a day and up to 1000 a day after a successful trade show. For example, Qosina may provide samples during a customer's R&D process, and only when the customer's product is approved for production will Qosina make its big sale. Quinn wanted a dynamic tool to support the company's atypical marketing and sales process. "Our processes, while not totally unique, weren't anything that could be supported with an out-of-the-box package," he says.

Quinn looked into customizing a hosted solution to accommodate the company's elongated sales cycle, but ultimately rejected the idea because the company already had the technology infrastructure in-house to support an on-premise system (Qosina has hosted its own Web site for 10 years). Quinn also wanted to retain power over the application and the data in it.

Security concerns often can be a sticking point for CRM customers. "Will my information be secure in a hosted environment? Will I have access to it? Who will own it? Will competitors be able to view my customers? These are important questions," says Wayne Latterell, president and founder of CRM consultancy Portico Solutions. "Imagine - would you put your company's financial data on the same server as that of your competitors?" Some customers will be satisfied with the hosted vendor's security measures, while others may not want to risk it.

"We wanted more control," Quinn says. "We didn't want to put our sales pipeline out there in a hosted environment with someone else controlling the servers." He had heard horror stories about ASPs in the dotcom era, such as some closing their doors without giving notice to clients, leaving them with no system and no way to re-create one, and others selling the customer data they hosted. While he knows that most hosted CRM providers today are much more reliable, Quinn says, "putting our proprietary information outside our firewall, spam filter, intrusion detection and virus-scanning software would be putting a great deal of faith in the hosting company. We may not have the same global resources as larger hosting companies, but at least internally we can hold someone accountable."

Quinn looked at several CRM systems including Siebel, Saleslogix and Goldmine, but ultimately chose Microsoft CRM so that he could more easily integrate it with the Great Plains ERP package Qosina had put in place the year before.

Page Break

Dialling into the Back Office

The level of integration required between a front-office CRM system and back-office systems is another factor to consider when choosing between hosted and on-premise CRM. On-demand CRM vendors are offering ever-more robust integration tools. But, says AMR's Bois, "integration is always going to be an issue with software-as-a-service because you don't own the application or have access to the source code". And more sophisticated real-time integration with back-office transactional systems isn't possible with on-demand CRM software - at least today. "There's movement in that direction, but they can only import flat files asynchronously in batches," says Bois. "Companies that need to do that kind of [real-time] integration are more likely to stick with licensed software."

"It's not that on-demand software can't integrate," says Greenberg of The 56 Group. "It's just that the integration tools in traditional on-premise software are better. The more complex the integration requirements, the better off you'll be with on-site software."

ResortCom's Marxer says the integration of the RightNow on-demand software with his back-office system is satisfactory, but not what he would call ideal. "RightNow delivered on all the integration points we needed and the performance was reasonable," he says. But to open up a customer incident report, which an employee does only when an incident cannot be resolved on the first contact, takes 10 seconds because of the back and forth on the back end. That's just fine for ResortCom's needs at the moment, but "there's some room for improvement", he says.

Mike Davis, CIO of Stewart Information Services, agrees. Like Slusar at SunGard, Davis had to think about the needs of the far-flung sales and customer-service organization supporting the $US2.2 billion title-insurance company he works for. With a requirement that the software eventually support at least 4000 and as many as 8000 users, Davis had the option of becoming one of Salesforce.com's largest customers to date (in fact, small pockets of people within the company had already started using Salesforce.com on their own). But unlike Slusar, Davis ultimately purchased a licence for Onyx's on-premise CRM product.

Davis wanted to tie the CRM to all of Stewart's "day-in and day-out" systems. "We needed the most flexibility we could get in integrating it," says Davis. "And [Onyx] seemed to have much easier ways [than Salesforce] to integrate our system with theirs and theirs with ours - three different levels of embedding and exposing the information. There were modules of code available to use within our systems to make it easier."

Salesforce.com also has ways to get data in and out of Stewart's applications, in an import-export fashion, Davis says. But it would have required users to manually initiate the imports and exports in a less-seamless fashion than he would have liked. "We might have been able to make it work," says Davis. "But it wouldn't be very efficient. And it wouldn't have made for a very good user experience."

Easy for You, Difficult for Me

One of the major selling points for on-demand CRM is its relative ease of implementation, particularly in contrast to the expensive and lengthy roll-outs that have plagued the traditional CRM customer.

Indeed, at Qosina, the Microsoft on-premise implementation took more than a year. And the biggest cost was consulting fees, which, at $US280,000, made up half of the implementation expenses. Davis of Stewart Information Services is just finishing his Onyx pilot (for six sales-force units), which also required the added expense of two full-time and two part-time consultants. It took seven months longer than expected because midway through the process, Davis discovered some additional functionality that would be needed for the regional sales offices.

"It's taken longer than I thought," he acknowledges. "The biggest hold-up just has to do with change in general," including getting users to adopt the systems and change the way they work. The integration that drove the decision to go with an on-premise product won't even happen until the next phase of the roll-out. "It's a long road," Davis adds.

For companies that can adjust to an on-demand CRM system out of the box, implementation takes less time. But it would be a mistake to assume that all hosted CRM implementations are quick and easy. In fact, most take time to roll out enterprise-wide and many require bringing in consultants to help out. "Some customers have the expectation that you flip a switch and you're done," says Bois. "But there are setup costs and training costs and ramp-up costs. There's getting the system customized to match the business context and then getting people to use the system. CRM implementations are still complex, even if they're delivered in an on-demand fashion."

Page Break

At SunGard, a smaller and more narrowly defined roll-out of Salesforce.com took a full year. "We had to do a fair amount of work lining up what we wanted to accomplish with what [Salesforce.com] could provide," Slusar explains. SunGard also had to bring in consultants to assist with the roll-out and accompanying change-management issues. Some of that effort was devoted to standardizing data definitions in order to preserve data integrity. Because the company had so customized the software's template to meet the needs of its global sales force, there was no way Salesforce.com's online training would work. Slusar rolled out the system in phases until 90 percent of her user base was on it by the spring of 2004. "We still could do more to drive standardized behaviour and usage that we haven't," Slusar says.

Looking forward, Slusar has created a team to oversee the Salesforce.com product over the long term. They will work to increase user adoption and also keep track of new offerings from the vendor as well as from "the mushrooming number of partners that have popped up all over claiming they have something that works well with Salesforce.com", says Slusar. (Applications offered by companies partnering with Salesforce range from Web-based HR recruiting and screening tools to call-centre scripting and software that helps doctors take notes on wireless handhelds.)

"The good news is there are a lot of things you can do with this system," Slusar says. "The bad news is you need people to do it. And I already have a full-time job."

Different Delivery Model, Same Risks

Marxer is sticking with the hosted solution, even though RightNow had a clause in its contract that ResortCom could buy a licence and take the software on-site if Marxer wasn't pleased with the results. Although he's had to hold off on upgrades to date, he's hoping they won't pose a problem when RightNow adapts a true Web-services integration platform.

In the meantime, the on-demand system has yielded some solid ROI. The self-service function has reduced customer e-mails by 40 percent, and workflow has improved between back-office and front-office functions, resulting in a 40 percent improvement in productivity. And ResortCom has been able to introduce some automated marketing functionality that Marxer predicts will boost its bottom line this year.

Marxer credits the progress he's made with his on-demand CRM solution not to the technology itself, but to the two years he spent preparing the organization for the changes that would be required - preparation that would have been the same regardless of which delivery model he chose. "Back then, we didn't have the money or maturity to do a full-blown implementation. [As a result], I was able to do all the prep work necessary on a cultural level way before we even put out the RFP," Marxer says.

It remains to be seen whether more complex on-demand CRM implementations ultimately succeed or suffer the same fate as many large-scale on-premise CRM projects have. "Because most of these are monthly investments and in the past were smaller deployments, we haven't heard about the big disasters yet," says Bois. "The larger implementations will be the ones to watch," he says.

"On-demand CRM may be a less-expensive risk, but it's just as big a risk," says Greenberg from The 56 Group. "If you haven't planned everything out, customers will get lost. People will get fired. You will fail."

Weigh In

SIDEBAR: Is CRM Dead?

By Allen Bonde

Alive or dead, CRM is vastly changed from the acronym we once thought we knew

With the acquisition of Siebel by Oracle, many of us are pondering the question: Does this mean that CRM as we know it is dead? As a long-time industry watcher and former analyst, I actually do look at this deal as a potential endpoint in the evolution of a model that has developed over the past 15 years or so. But I also see it as further evidence of the sea change taking place in the overall enterprise applications market, which despite challenges and the potential changing of the "old guard", is in fact undergoing a bit of a renaissance - especially when it comes to bringing powerful new capabilities to the masses of business users and empowering customers to better serve themselves.

Customer relationship management or CRM as a model has its roots in three primary areas: call centre systems, help desk applications and sales-force automation, or what some have called the "front-office functions". In the mid-1990s several platform providers like Siebel and Clarify (now Amdocs) emerged, driven primarily by acquisitions, to offer consolidated functionality across the entire front office, while the "back-office" providers like SAP and Oracle generally remained focused on areas like finance, supply chain management and as it emerged, e-business.

In a way, the consolidation of front-office and back-office functionality under one umbrella - like we see with the Oracle-Siebel deal and saw before that with the PeopleSoft-Vantive deal - has been a long time coming. The benefit of having one database and common set of end-user tools is attractive. Plus, the return on traditional, stand-alone CRM investments has been mixed at best, especially when it comes to large-scale deployments. I know of several global organizations that have spent more than $100 million on CRM projects and are still uncertain what real value they have received!

It's (Still) About the Customer

Despite its name, one can argue that the greatest shortcoming of CRM is that it never really was about directly helping customers. Solutions were sold to executives running call centres or sales organizations as a way to wring out inefficiency, force standardized processes and gain better insight into the state of the business. In particular, what most CRM and CTI systems provided was a way to track customers, route and facilitate inbound communications and report on the progress of various marketing, sales or support activities.

But what these solutions generally did not address was the need to help organizations resolve customer problems, answer their questions faster or help customers solve their own problems. For this reason, we have seen a slow but steady shift in focus and investment from automating core internal front-office functions to streamlining edge processes like online customer support, product returns or account management.

In parallel, there has been a recent wave of innovation powered by Internet standards, open source software and on-demand delivery models, as well as a renewed interest in areas like knowledge management and what some are calling service resolution management or SRM. As defined by leaders in this sector like Knova Software, SRM aims to improve access to corporate knowledge by breaking down silos, simplify the authoring and capture of new content and provide more consistent answers across all sales and service channels.

Other innovators who are filling the gaps inherent in "old-school CRM" include e-commerce and personalization pioneer ATG; RightNow with its innovative on-demand customer support and self-service offerings; e-billing and online account management specialist Netonomy; Genesys and Talisma with their customer interaction management solution platforms; and content optimization specialist SafeHarbor. At the same time, several vertical solution providers like Astute Solutions and Chordiant have created next-generation applications which fill industry-specific requirements.

A New Model for CRM v.2

What is the future of CRM - or CRM v.2 if we decide that CRM v.1 is, in fact, dead? First, there must be a core driven by business rules and even knowledge management, rather than just a database. Second, solutions must address all modes of interaction, whether with an agent or salesperson, on the Web via self-service, or peer-to-peer via user forums and other collaboration techniques. Third, solutions must be adaptive, by applying analytics and personalization approaches, so that organizations can anticipate customer needs, proactively push out solutions, recommendations or offers based on who the user is, their skill level, what their preferences are, and so on.

More generally, the on-demand delivery model appears to be here to stay, although we feel that all deployment options should ideally be supported. The use of open source and developer source-based components such as those from Jive Software are also gaining momentum, especially for multi-channel "edge" functions like customer forums or enterprise instant messaging.

For the design centre, we look for CRM v.2 to be simpler to use, more open and more adaptive. It also must be inherently multi-channel, as Gartner's customer interaction hub model suggests. The customer adaptive solutions theme announced at Siebel CustomerWorld also seems on the right track. And there is the argument being made by Greg Gianforte at RightNow that on-demand delivery coupled with open source infrastructure may be the most efficient way to bring these types of applications to market, a view that has a lot of merit.

So, while CRM as a model is continuing to evolve, as a market it is definitely entering a phase where "big-bang" deployments are likely to be the exception, and ways to more efficiently reach underserved users and improve responsiveness via add-ons like user forums, a self-service knowledge base or customer analytics become the focus. Solutions will also need to be integrated, if not as part of one platform, at least in terms of common standards, and a common focus on the customer rather than only customer processes. The future of CRM as a viable approach and market depends on it!

Allen Bonde is the senior vice president of strategy & marketing at eVergance, a management consulting and systems integration company focused on CRM optimization and Web self-service