self-service - DB Kay & Associates

Metrics Myth 10: Contact Deflection = Clickthroughs Times the X-Factor

14 May

Industry surveys ask. Colleagues ask. IT asks. Executives ask, incessantly.

“What’s your contact deflection rate?”

Let’s leave aside for the moment that contact deflection isn’t the primary reason to expose your knowledgebase to your customers. Let’s also leave aside the fact that it’s both practically and philosophically difficult to count things that didn’t happen. (Do I get credit for all the times I wasn’t snippy to my family and colleagues? No? But I counted 39 instances of snip-avoidance last week! I mean, if I hadn’t been so nice, I would have been a real jerk.)

So, like it or not, we have to come up with an answer for how many contacts we’ve avoided. As an aside, I don’t recommend that you characterize this as a rate—that is, a proportion of something else. People ask for this, but in my experience most don’t really know how to think about it that way, and there also isn’t universal agreement about what to divide it by. Saying “We had 10,000 self-service sessions this week that deflected approximately 500 cases” is clear enough; saying “we had 68% contact deflection” seems less obvious.

I’ve written elsewhere about techniques for estimating contact deflection; if you’re interested, please look at this newsletter article, and I won’t wave mathematical formulas at everyone else reading this blog.

What I do want to say is, it’s not straightforward. And I cringe when I hear that there’s some magical “X Factor” that you can multiply by page views or web sessions to calculate deflection.

For one thing, contact deflection is lower than many people expect, because the majority of customers using self-service, even entitled enterprise customers, wouldn’t have opened a case. As charming as we all are, customers would usually rather not talk with Support if they can avoid it.

But whatever your customer’s propensity to open a case after an unsuccessful self-service interaction (or page view, or Google search, or in-product link, or…what were we multiplying again?), you have to acknowledge that it’s going to vary from customer segment to customer segment, and product to product, and perhaps by geography, even within your own company. Certainly, your customer behavior around opening cases after self-service will be different from that of other companies, even in your industry.

So the idea that there’s a single multiplicand that can give you self-service contact deflection is, unfortunately, just another myth. If you must have an answer, calculate escalation rates and success rates as best you can, and let everyone know you’re estimating. Please don’t say that you’ve multiplied your web traffic by an “industry best practice” constant.

Little Data

By: David Kay 0 Comments

20 Dec

You can’t escape hearing about Big Data right now—it has replaced “cloud” as the current hot buzzword in tech. The idea is pretty simple: it’s easy to gather huge volumes of data from Internet clickstreams, scientific instruments, roadway sensors, and all the other collectors of digital information in the world…but it’s not always easy to process it. Big Data techniques are designed to rapidly sort through these piles of bits and bytes in order to gain meaningful, actionable business insights…such as the fact that you like your friends’ cat videos on Facebook.

So, the technical definition of Big Data is all about how to do correlation and trend-spotting in near real time on petabytes and exabytes of data. But in the popular imagination, I think it has come to mean something a bit broader. Big data evokes a sort of automated Big Brother—a bank of servers tirelessly poring through tweets, searches, mobile location data, and YouTube comments. Except, unlike Big Brother, most of this surveillance is designed to figure out how to send you the perfect marketing pitch, all without human intervention.

It’s exciting technology, but it’s a little bit creepy. In our business, I think we need a complementary approach. We need to get really personal with our data. Let’s re-discover Little Data.

Most service and support leaders’ questions are hard to answer just by crunching the numbers.

What one product change would help customers the most?
How much is knowledge shortening resolution times?
How effectively is self-service helping customers?
What’s the loyalty impact of a particular support initiative?
How well are we collaborating?
How many contacts are we deflecting, and are they the right ones?
What’s the ROI of knowledge?

These Little Data questions require judgment as well as correlation coefficients; nuance as well as regression analysis. They require estimations, surveys, interviews, and other less-than-perfectly-precise and human-intensive inputs.

But, as measurement guru Dean Spitzer says, it’s better to measure the important things imprecisely, than irrelevant things with perfect accuracy.

I’m excited to have just wrapped up one of these projects, focused on voice of the customer. We have lots of Excel sheets, and StatPlus got a workout. But the interesting part was thinking about the right questions to ask. To me, the most satisfying results were the stories that emerged from the numbers…and the discussions and innovation triggered by those stories.

Something we hear all the time is, “we have all this data, but we haven’t had a chance to really dig in to it.” Maybe it’s time for a Little Data project.

ps – registration is now open for our next KCS Foundations Workshop. Anyone up for February in California?

Keeping Two Sets of Books

By: David Kay 2 Comments

23 May

Photo credit: iStockphoto.com

I learned a great best practice for organizational change this week at the Consortium for Service Innovation’s executive summit in beautiful Chatham, Cape Cod: Keep two sets of books! No, not in the bad way—not like Enron—but to enable change.

Several industry leaders are keeping two sets of books to track performance on their current organizational obligations, while setting the stage for a new way of thinking about their work.

For example, let’s say you’re in charge of a self-service initiative. Readers of this blog know that self-service is a customer engagement strategy, not a customer deflection strategy…but do you want to have that conversation every week with the CFO?

A better way is to keep books both in the organization’s currency of the day (deflected contacts, avoided operational expense) and in your aspirational currency (loyalty, feature use, and repurchase, for example).

This is an easier conversation: we start off by showing our stakeholders and funders how we’re doing the things they care about, then we earn the right to have a conversation about why they should care about the things we care about, too.

Also, emerging measures are usually imprecise and somewhat subjective, as Dean Spitzer tells us. In contrast, measures in our “first” set of books tend to be better vetted, and more familiar to the organization. These more conventional measures give us breathing room to gain experience with our new measures without subjecting them to the third degree.

Without aspirational measures, we can’t refocus the organization on new and innovative sources of value. Without our conventional measures, we’ll spend all our time arguing and justifying.

As soon as your new measures take hold, you get to institutionalize them in the first set of books—and you have the great opportunity of creating even more innovating measures, keeping the change moving forward.

So, keep two sets of books. (Just please, don’t call it that when you’re presenting to the CFO.)

A special tip of the hat to the support executives who shared their “double entry” stories and successes, especially Medi Goker of Oracle; Marco Bill-Peter of Red Hat; Steve Young and Janet Ramey of Cisco. The good ideas in this post are theirs; misrepresentations of their points of view are mine alone.

How Self-Service Fails

By: David Kay 0 Comments

17 Feb

Photo credit: Arrr! Via Flickr. Creative Commons Attribution license.

Often, the first step in making something work really well is to figure out how it can fail…and then making sure it doesn’t. In the case of self-service, most studies show that customers are successful less than half of the time (sometimes far less.) So clearly, there’s plenty of failure to eliminate!

Here are some ways we’ve seen self-service fail, from our engagements with clients and from our own personal experience. How many of these barriers to success might frustrate your customers? And how many of them can you eliminate or reduce?

The answer to my question isn’t here. (This often speaks to traditional, slow-moving knowledge management practices—consider KCS.)
Or, if it’s here, I can’t find it. (Is knowledge written from the customer’s point of view, using their words and context? Do you have a competent search engine that returns relevant content from any source?)
I found the answer, but I don’t like it very much. I’ll try calling. (Are you more flexible on polices, for example, warranty returns, if your customer calls? You teach people how to treat you…)
My eyes are glazing over. (Is your portal neat, clean, and streamlined? Just because someone created a tool doesn’t mean that it deserves a link on your support home page.)
Huh…what? (Do you require your customers to learn your organization and your jargon?)
Don’t blow sunshine up my skirt. (Look on the bright side of life, but don’t try to rewrite a break-fix answer as a how-to. Self-service is not the place to spin the facts–that will just send users to find someone they do trust.)
I bet you know, but you aren’t telling me. (This is what web maven Steve Krug means when he says a website should “be a mensch.” If people think you know about a problem but aren’t talking about it, they lose faith in you and your product or service.)
I wanted an answer, not a dissertation. (Any words that don’t take customers one step closer to a resolution should be eliminated. Link to more complete background explanations—right now, let’s make it “just the facts, ma’am.”)
The answer isn’t showing up on Google. (Is there any way you can make at least some of your content available without a login? And have you invested in Search Engine Optimization (SEO?) Increasingly, whether we like it or not, search engines are the self-service portal of choice.)

What self-service problems are you trying to avoid?

Resource Roundup

By: David Kay 2 Comments

19 Jan

Sometimes we find so many good things that we can’t help but stand back and let others speak. This is one of those weeks. So, without delay, some links we love.

1. In a recent installment of his Eye on Service blog, John Ragsdale tells a sad story that’s all the sadder because it’s so familiar. Running into technical problems when deploying two Christmas presents, he laments

“I ran into the same problem I have any time I attempt self-help–my problem doesn’t exist in the knowledgebase. It is beyond frustrating. You encounter a problem that many new customers are likely to run into, and there is nothing online to address it. Usually, you can find hundreds–or even thousands–of conversations in a forum about the problem, yet the knowledgebase contains not a single reference to the issue.” (Read more…)

Isn’t it time that we all have a single knowledgebase for internal and external use? And isn’t it time that we use KCS, or some way of making sure it’s up-to-date? This is old news to readers of the DB Kay blog, I trust, but our industry still has a ways to go.

2. Steve Krug, author of my favorite web usability book Don’t Make Me Think, promised that he’d follow up with a more detailed book on lightweight, DIY usability testing. Nine years later, here it is: Rocket Surgery Made Easy. If you have anything to do with a web self-service site, or any other web experience, please buy and read these books immediately. Oh, and while I’m recommending books about the web, check out the wonderful Letting Go of the Words by Ginny Redish, who provides practical and interesting coverage of the KCS topic “complete thoughts, not complete sentences.”

3. Finally, Social Success, our paper on social support, has been published by the Consortium for Service Innovation. This summarizes several meetings’ worth of insights from Cisco, Yahoo!, Lithium, Salesforce, EMC, and many other Consortium team members. Read the paper…

Communicating KM to Customers: IBM does it right

By: David Kay 1 Comment

4 Jan

Our friends at IBM Information Management just shared a wonderful video they’re using to communicate about their KCS practices to their customer base. Have a look on YouTube.

Some things I love about this video:

It’s short. It gets all the big ideas across in three minutes. As someone who routinely talks about KCS for three days, I admire this. And, I feel like I got a behind-the-scenes view of the new way support happens at IBM.
It’s concrete. There are no lofty principles or busy “marketecture” diagrams here; it’s all show-don’t-tell. Communicating this way looks easy, but it’s not.
It has the right style. It’s professional, but not at all slick. It looks IBM official, but still very human. Wouldn’t you like to have Alessandra helping you with your support issue? I would.
It explains WIIFM. It keeps highlighting what the customer will experience, and why that’s a good thing for them.
It has a call to action. By the end of the video, IBM has convinced me I ought to try the support portal first…oh, and put feedback on their articles. too.

Does your communication plan reach out to customers? Here’s a great way to do it. Nice work, IBM! (HT Mary Ellen Coleman).

ps – speaking of videos, our KCS In Five Minutes just went over 1000 views. I know that’s technically not “viral,” but if feels pretty good. Check it out and pass it on if you like it!