Going Steady

Article Featured Image

Billy Joel was right about relationships—not romantic ones, of course, as his own storied track record disqualifies him on that front. But when applied to vendor relationships, nearly every platitude from his 1983 song “Tell Her About It” holds true in the space.

This is probably because hosting a contact center speech solution through a vendor agreement is a lot like a marriage. In the process of partnering, a customer is entrusting a core of its customer-base interaction to a third party. If the relationship works, then it can have real benefits, but, as Joel would also tell us, it will only work in an environment of open communication and willingness to partner, which isn’t all that easy. To facilitate that, a project needs clear goals and standards with appropriate incentives and penalties—in short, a strong service-level agreement (SLA).

But first it might be helpful to clear up a few big misconceptions that vendors and analysts say customers often bring to a hosted relationship.

For starters, know that hosted isn’t always cheaper. In fact, it can be more expensive in the long run. When customers talk to vendors, many times they’re dazzled by the promise of an almost immediate return on investment, unlike an on-premises system that requires a big upfront capital investment. But, the glittering figures one sees in initial bids are usually tracked out no more than a handful of years. Many systems, says Eric Tamblyn, vice president of marketing at Genesys Telecommunications Laboratories, are in place much longer—as long as a decade.

“A hosted provider has all the same costs a customer has,” he explains. “In fact, not only do they have the same costs, but they have to turn a profit, so they have to include their margin to make money.” 

Tamblyn adds that hosted providers can absorb, normalize, and minimize some of the cost since they’re working at a larger, devoted scale, but like anyone else they have the same contact center overhead of servers, facilities, and staff, and those costs catch up with the initial on-premises costs. Within two or three years, hosted “blows through the cost” of on-premises, Tamblyn argues.

All that said, for some businesses, like florists and others with intense seasonal spikes in traffic, a hosted solution might prove to have a much higher ROI when compared with an on-premises solution. 

For the rest, the advantage of a hosted, speech-enabled contact center is that they are shielded from risk. By paying someone else to manage their systems, they mitigate the risk of getting the system wrong or getting up and running late. This avoids a lot of potential money wastes. They’re given access to expertise—collectively thousands of years’ worth in some firms, Tamblyn says—which is often hard to come by.

The second misconception is that hosted always means up-to-date. Many people wrongly assume they will automatically have the latest release of a platform from their benevolent and giving vendor. Not always so. On most multitenancy platforms—that is, hosted servers that the vendor uses for a number of different domains—customers benefit from automatic system upgrades. On multitenancy platforms, the vendor can upgrade several of its clients with one software refresh and with hardware for one system. On dedicated servers, though, upgrades have to be done on an individual level, and then only with the explicit approval or at the behest of the client, who, in turn, pays for them directly.

In an SLA with a vendor, this is usually signaled by a clause that says something to the effect of: “We do not guarantee that we will update your software during X period.” If you see that, then you can probably bet you won’t be getting the latest updates unless you ask for them.

A potential hosted customer might also want to ask its vendor-to-be which version of its given platform it’s running. In fact, some value-added vendors are likely to be a few versions behind.

The takeaway here is that even though a company might be outsourcing to ostensibly more capable hands with the technical know-how to put a system together, those responsible for the outsourcing need to do research and make sure they understand what they’re getting into. This is just as important as the relationship develops too.

A Little Shy

Scott Manghillis, director of hosted contact center solutions at Convergys, admits he sometimes comes across customers who have no clue about what they need. “When they walk in the door and you start asking them a question, they are like, ‘I don’t know. Here’s the number to our current system. Go to town. Come back with something. I don’t want to deal with it.’” 

Conversely, other customers might hold back out of fear of exposing too much of their core business strategies to their partners. “One of the things we see quite a bit is that customers have done a lot of analysis and back-office research and have exposed very little to vendors,” Genesys’ Tamblyn says. 

He describes a customer who tries to do as much as possible in-house, not talking to the hosting vendor about business ends and goals, but just trying to pick up the technical expertise and manage it. Such customers come to him with a “checklist of 300 things”—usually focused around functionality. If enough items in the checklist line up for the right price, the customer is happy. The problem with this approach, according to Tamblyn, is that it leads to breakdowns in the vendor-customer relationship and, ultimately, the automated contact center domain.

“Six years ago I got invited to a conference call.… I realized it was going to be a bad call when I saw that the bridge for the conference call was going to be ‘hard ass.’ So I got on this bridge and listened to the [chief information officer] tear into us and call us every name in the book—all that stuff. I understood the issues and the concerns, but what he said was ‘By not having this system online I am losing a million dollars a day.’”

Tamblyn recalls being completely blown away. “We spent two months with this customer chiseling every technical consideration possible about what our platform did or didn’t do, but we never talked—not once—about the business impact of what we were trying to do. We asked over and over, but they gave us very little input over when it was going to go online, and what it was going to do that was so critical in terms of what they had in place already.”

Genesys, it should be noted, licenses its platform through third parties making a direct, white-label offering. Tamblyn’s story, though, is illustrative. Withholding core elements of business aims can be tantamount to sabotage in a hosted partnership. 

As noted earlier, the cost savings from a hosted arrangement only last a couple of years. One of the major advantages to hosting a contact center is access to expertise. When you don’t make a vendor part of that strategy by informing it of business goals and aims, then you effectively limit its expertise to that of a pure technician and forgo its years of practical business experience; the “who” proposition of hosting seems far less useful. 

Job Insecurity

On the other end of the spectrum, some vendors complain that micromanagement can interfere with the process of building out a system. “If you [micromanage] and I have to explain everything that I’m doing, then the explanation can sometimes supersede the time spent deploying,” Convergys’ Manghillis explains.

That kind of resistance often comes from individuals worried about being automated out of a job—often folks in IT  departments who are skeptical of hosted altogether. He says he also sees micromanaging tendencies among management, but there the pushback is often a question of natural temperament. This is often borne of intense concern around ROI and cost containment.

Manghillis says when he comes across micromanagement from the executive levels of his clients, he has to “deal with that and move on.” 

“A lot of times we’ll throw an engagement manager in just to deal with that micromanaging mentality, to talk while other people are doing,” he adds—which means a greater resource allocation and probably higher overall costs across the hosted space.

The difficulty in both of these stances toward vendors—secrecy around business plans and micromanagement—stems from the same problem: a lack of trust in a vendor. That trust is essential to forging a hosted solution relationship. If a customer absolutely can’t have one, then it probably shouldn’t be hosting a contact center altogether. That said, trust doesn’t exist in a vacuum. The SLA is a big item that customers have in their toolkits for defining their demands and making sure their vendors deliver.

The general consensus among vendors and analysts is that users ought to largely define SLAs around specific business objectives and corresponding metrics. Most look to “customer satisfaction” as a gauge of success. Different vendors have different ways of calculating it, though. Sheila McGee-Smith, principal analyst, president, and founder of McGee-Smith Analytics, says the best way is to look to increasing minutes.

Getting Comfortable

“Especially with speech, there is a learning curve on the caller side,” she says. “As people get more comfortable with the solution, they should opt to use it more often. One of the markers for me, as a provider, should be that whatever the initial take rate is, it should increase over time. With that increase, I should be in the position to go in and expand the deployment. You start small. You start with trials. Success should breed success.”

None of the vendors interviewed for this article mentioned availability as a metric for success. They, as well as the analysts, think uptime should be a “given” in an SLA. Customers ought to demand near 100 percent availability, but that should by no means be the focus of an agreement or a barometer of how a customer defines the mission of a hosted solution.

Once the core mission of an SLA is defined, one of the biggest remaining challenges in its crafting is finding the right incentive structure to ensure a vendor holds up its end of the deal. 

Some customers feel a hard penalty structure that hits a vendor hard for network outages is the way to go. They might hold the vendor to a five-nines agreement (that is, one that requires 99.999 percent availability) and punish a vendor with heavy fines (as high as the return of an entire month’s revenue) for failing to meet that metric as a way of holding a vendor’s feet to the fire. 

This approach, however, might be too Draconian and could end up ruining the client’s own domain in the end, suggests Laura Marino, senior director of product management for Nuance OnDemand.

“Five-nines means that you get 23 seconds of downtime,” Marino says. “It’s a very strict thing. But let’s say the vendor has an issue, and the platform is down for more than 23 seconds. Right away the vendor is going to say, ‘I’ve lost all the revenue for this month,’ so there’s no incentive to fix the problem right away. They’ve already lost their money.”

Those slim 23 seconds could turn into days or weeks of delays and downtime as the vendor looks to its other, still-paying clients to make up the difference. 

Instead, incentive structures have to be proportional to the lapse, Marino says. If a vendor’s platform goes down for only a short while, then the revenue return ought to be similarly small and grow as that time extends or the problem grows more serious, thereby incenting the vendor to quickly fix any problems.

“At the end of the day, a customer will never win by getting paid back by the vendor,” Marino adds. “No matter how steep the penalties, it never compensates for the negative impact of having affected your end users.”

Making Minimums

On the customer side, enterprises ought to be aware of certain penalties as well. Chief among them are minimum traffic requirements. Many vendors will include these to ensure profitability for themselves for the duration of the contract, but particularly with longer-term deals, this can become onerous for hosted customers. 

Two years ago, the world market at large was expanding vigorously—too vigorously in the U.S., we now know. If an enterprise entered into a multiyear agreement in that environment, with certain expectations of incoming revenue and traffic, then it would find itself in a sore position now. That company would be paying its partner against minimums that might have proved higher than it could steadily deliver.

In light of this, McGee-Smith advises businesses entering hosted agreements to take a broad view. “It might make more sense to go with a lower commitment at a slightly higher rate than to commit to be paying against traffic you can’t deliver,” she says.

If an enterprise is trapped in a minimum it can’t meet, then it is left with few recourses. It could try to break its SLA, but this often incurs heavy penalties in disconnect fees, some of which could be the equivalent of the minimum for the term of the agreement. Obviously, in such a case it wouldn’t be worthwhile to leave a vendor because the payment would end up being the same. 

However, often a customer can negotiate with a vendor to get it to lower the fee to less than the minimum over the whole term. McGee-Smith says a disconnect fee worth the minimum of the contact can be a vendor’s opening gambit in negotiations. It’s the vendor’s way of protecting itself against rivals or the strange caprices of its customers; it locks customers—or at least their dollars—into the service during the entire contract.  

McGee-Smith also cautions that one shouldn’t expect to see the fee go away altogether. It wouldn’t be fair. Rather, she says the fee just shouldn’t generally “be as rich as a carrier or provider would like to see it.”

It’s a difficult give-and-take to manage, but customers have to be very cautious and forward-thinking when they draft an SLA. It’s no light charge. In many cases,  these are documents that outline the foundations of a relationship that could span a decade or longer. Before an enterprise signs one, it has to have clearly defined not only its current business goals, but what those goals are likely to be at the end of the contract’s term as well. 

Shortening the courtship

Ten years ago, many hosted solutions required a lengthy vendor approval process to put updates in place. A customer would basically have to call its vendor, tell it what it wanted, and then wait to move through the chain of verification at the vendor’s end. This was primarily because vendors were looking to protect themselves from harmful changes that endangered their solutions or, worse, brought down their entire networks.

Many vendors still take an approach that requires approval, but as technology has advanced, others are offering Web-based applications that allow for more direct user control and immediate access to analytics tools. For example, in some systems, if a user wants to make a routing change, she can do it herself. The change would still await vendor approval, but the change could be made directly by the user, allowing enterprises to be a little more limber. How much control you get on the back end is definitely worth consideration. 

Monogamy’s not always the answer

Sometimes enterprises don’t shop competitively for their hosted providers. If they have a relationship with one vendor already—say, one that’s doing their diversified routing—then they might be more inclined to just go with that same vendor again to simplify the vendor relationship. This isn’t always the best tack though, says Sheila McGee-Smith, principal analyst, president, and founder of McGee-Smith Analytics.

“Step outside that traditional mindset to say there might be other people out there who give me more choices, more control, and maybe I, as the contact center manager, may have to push back on IT a little bit more to get the solution that’s better for me, but maybe it’s worth it in the long run,” she suggests. “You might get a richer application.”

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues