Your call may be recorded …

Today I spent a lot of time chatting with IVRs.  Two of those experiences stuck out because they were positive IVR experiences. Two in a row.  Somebody … no, two people,  who are in control … understand that IVRs are a real part of the customer experience and the content can be designed.

One case sticks out because the design  minimized the message length without compromising the clarity of content or what the caller should do. That goal is not surprising. It saves the organization money. Every second on the line counts. What was interesting was the way they went about minimizing the interaction duration.

The other stuck out for the opposite reason. Someone made a conscious decision to lengthen the industry standard message. The longer message was brilliantly designed to buy indulgence(s), just in case the customer service  wasn’t as stellar as promised.

In the short-as-possible message the word strings for the menu options were not repeated slavishly for each item, as often happens. Instead, some words are omitted. But the omission was  thoughtful. (If you read it out loud and you will get a better sense of the effect.)

Look again … here it is showing the deleted words.

Notice that the sequence starts complete (“For A, say A or press 1″, gets short (E, 5 …) and cycles back to the full structure. There are quite a few options. Over time people (who are not paying close attention, really) forget. But the design protects them from their dual tasking tendencies. At least it did me.  Subtle. Clear. And this design saves several seconds (or more) per call. That multiplies up quickly in IVR ROI-land.

In the the case of the longer-than-absolutely-necessary message, I needed to get to a customer service human. No way around it. But the first thing offered was not a long Pick-1-for …. menu. And it was also not the standard:

Your call may be recorded or monitored.

Instead, the intro message was:

Here at the XX Platinum Customer Care Center we are in a an ongoing effort to improve our quality customer service.  Your call may be taped or monitored  for the coaching and development of our associates….

This message is longer  than the standard message. That’s an added cost.  (Accounting, perhaps, for the speeded speech?) … But the message has an interesting sub-dialog that likely more than returns the additional cost.

I suspect that it is a bit of an inoculation effects. (Like shots … If you are exposed to something in a low, safe level, you are less likely to respond adversely when it appears at a potentially toxic level.

The message embedded in the message is a very low does of  “Sometimes we get it wrong.” Its conveyed as “We know that sometimes we get it wrong. So we are actively coaching and providing feedback to the associates. We are working on making it better.”

In the worst case, where customer service is not as customer oriented as it might be, that little bit of inoculation serves as an implicit,  proactive apology. It likely buys the customer service agent a bit of patience. And by extension, it probably minimizes escalations. Which would translate to ROI.

In the best case, where the customer service agent is polite, professional and effective (my case, btw), you are left thinking, “Hmmm. They are doing pretty well, all things considered. And they are still monitoring for improvement opportunities. This is a company that gets it.”

Its a Win-Win. For a few cents a call.

I suspect, given the quality of the strings, both companies conducted live A|B testing of variations on the messages before they went to full implementation. If not,  they have IVR designers with very good design intuition. For the rest of the organizations out there with IVRs, it is  fairly trivial to design and run limited sampling studies that demonstrate unequivocally that customers can follow the streamlined message effectively (in the first case) and that the additional message content reduces escalation incidences. And that testing woudl be the basis for a concrete ROI report.

We may never know. But in the mean time, it provides hope that other organizations will also realize that deploying evidence-driven designs in their their IVRs could save a bomb … and even (possibly) make customers more friendly.

Mobile Persuasion … link to slides and talk

Instead of writing … for the last bit I’ve been thinking about (and doing) this contribution to the UCL  Mobile Persuasion meeting.

http://handheldusability.wordpress.com/

Now back to the regularly scheduled analytics …

Undo unto others….Twitter’s imperfect study of reciprocity and (un)requited love

Every once in awhile pop culture gives psychologists a gift.

Survivor, for instance, was a gift. The Apprentice, Parking Wars and now Smile, you’re under arrest… also gifts. In fact, the entire genre of reality shows is s a window for observing influence strategies and obedience-to-authority in a semi-controlled setting.  Authority. Social Proof. Ingroup-outgroup distinctions.  Reciprocity. They are all there for the watching.

And now we have a new gift: Twitter (www.twitter.com). Twitter is the emerging social network starlette, a microblogging tool. Its a bit like Facebook: You sign up to follow people you like and/or who share your interests and you exchange information. Except that on Twitter, that information is limited to (hopefully) profound and pithy statements of 140 characters or less. Think of it as exchanging headlines or pointers.

Twitter is a rich space to study reciprocity. Reciprocity is basically our drive to return favors. Reciprocity is why, when someone makes you dinner, you feel obliged to make them dinner back.  And why when someone follows you in Twitter, you feel a little guilty if you don’t follow back. At least the first few hundred times.

Osen Komura Twitter Stats

For psychologists,  Twitter could be a space to answer questions that we can’t necessarily ask in a university research lab. Internal Research Review Boards are reasonably twitchy about internet research and privacy, after all. The scale and reach of Twitter is big, too.  And, at its essence,  Twitter is reciprocity in its most kneejerk form.

Social psychologists should take note: Armchair psychologists are already running quasi-studies ….

For instance, Stefan Tanse, a student in Romania, created the Twitter account Osen Komura  (www.twitter.com/osen) to explore whether people automatically “followback” strangers who follow them. He posted a few benign messages, signed up to follow 40,ooo other people and watched what happened. When he quit watching sometime in April 08, 12% of the people that he followed had signed up to follow him back. That’s a pretty good rate for a feed with no real content. And his following rate is still growing. Today (30Dec08) his followback rate is 24%. For sure, this exercise is well known. Myriads of bloggers have highlighted the “Who is Osen Komura and why is he following me?” phenomena.   But if people read the blogs, they’d know  … oh never mind.

Nantela (a biologist from Montreal) expanded the experiment in two ways.  First, he wanted to see if people really do just click “follow” reflexively or if they actually look at the persons feed before they decide to follow it?  To find out, he created Ru4real and repeat posted a single message to Twitter: “DO NOT FOLLOW THIS ACCOUNT. It is an experiment to see how many people read the pages of the people they follow. You are a tool.” And signed up to follow 6,ooo people. The blind followback rate was about 10%.

But here’s where it gets interesting. Nantela also wanted to see if reciprocity applies for negative events, too. That is, if I unfollow you, will you unfollow me? (Maybe we call this revenge?)  So he unfollowed everybody he had been following. If reciprocity applies for negative events, then we  expect  the people that Nantala unfollowed will  unfollow him, too.

Nantela’s observations suggest that we do also reflexively return for “negative favors”. Unfollowing happens. Ru4real lost two followers just yesterday.  Nantela’s “results” suggest that retaliation for unfollowing happens at a slower and lower rate than for following. Be careful, though. That’s likey just an artifact  of how Twitter works. Today,  Twitter’s default alerts you when someone signs up to follow you. But it doesn’t do that that when someone UNfollows you. So, while there is an clear and explict “follow” event to respond to, there is no such “unfollow” event to reciprocate. So Twitter needs to change a little before we can learn whether for if revenge is as reflexive as it is sweet.  Note: Some Twitterers opt, through add-on programs, to be notified when someone unfollows. I suspect those people are reflected in the  droppoff cliff in the followers graph above.

Maybe Twitter will change. I hope it does, actually. Running this “experiment” for real could extend and refine our understanding of reciprocity and and interpersonal behaviors within both Twitter and within the social networking universe.