Good robots vs bad robots: two cases of customer service automation
Good robots versus bad robots: in customer service automation, there are winners and losers. There is good customer service, and there is bad customer service. What makes for a good automated experience versus an evil one?
Cross-posted to Talk It Up
I write a lot about customer service here, especially as it relates to the call center. And I get a little tired of the all-too-fluffy blog posts on customer service that give vapid advice (Smile when you talk!) or simply recount a bad experience to illustrate the fact that, in fact, bad customer service exists.
We’re aware that bad customer service exists and that good customer service isn’t as easy as it looks. And one of the controversial aspects of customer service is the trend toward automation. Some users proclaim their hatred loudly and flock to sites such as GetHuman. Others… well, no one really sings the praises of a well-designed IVR or email ‘bot, because when they are beautifully designed, they simply work as promised and let the user move on to focusing on more important tasks.
Over the last week, I had two experiences with automation in customer service. With one, the automation was evil: involuntary to begin with, and the organization prevaricated, blamed the user and required several inquiries and escalations to complete a basic process. With the other, the automation was so simple as to be practically invisible, anticipating the user’s every need. What can we learn from each one?
Automation #1: Email list auto-add (Evil)
With the first experience, I was added to an email list for night club events taking place in the D.C. area. I’m unsure as to how I got on the list, since I didn’t subscribe and the email in question is used almost exclusively for listener feedback from a podcast I produce. It’s not even really my email; it’s the show’s email address. At any rate, I sought an unsubscribe button, since (a) I live 3,000 miles from D.C. (b) I don’t frequent nightclubs and (c) I generally don’t drink cocktails or spirits.
The first thing I did was seek an unsubscribe option on the offending emails. No such luck. But I was given the prominent option to forward to a friend.
So, grumpy at the extra effort, I was forced to email the organization to report the spam and to ask to be taken off the organization’s list. What followed was a series of ill-mannered prevarications designed to make it extremely difficult for the user to be removed from the list. The organization responded, insisting that I use a nonexistant unsubscribe button. When I pointed out their newsletters not only do not have an unsubscribe button but that I’d never signed up in the first place, they insisted I had signed up for the list (!!) and that they were unable to remove me. Only robots can add and remove customers? Evil!
My response pointing out the lack of an Unsubscribe option:
I reported the behavior to the organization’s email service, as most services have regulations regarding spam and automated subscriptions. Also, it’s unlikely that the organization was unable to manage its user list and delete users on request, and it did turn out to be untrue, as the representative at iContact was kind enough to point out: “As for adding and removing contacts, the sender does have those capabilities. If an iContact customer is claiming they are unable to perform this task please let us know so we can look into the situation.”
After some discussion about the lack of an Unsubscribe button, the issue was resolved by the email service provider–but never by the organization itself.
What went wrong:
- Automating without consent. It is absolutely best practices not to subscribe a user to a list without explicit consent. Even if I had attended an event once, that does not equate to consenting to being added to an email list. Likewise, every email list message must provide a clear unsubscribe option; hiding behind “Update Email Address” is no excuse or substitute. And let’s face it; it’s a little sleazy.
- Arguing with the customer. I’m not a believer that the customer is always right; we’re wrong all the time. But arguing is fruitless, especially with regards to simple transactions that occur every day. Best to make a polite apology and remedy the situation.
- Blaming the customer. Claiming that the customer’s situation is her own fault is rarely helpful, even on the occasions when it is true. (Although in this case, it wasn’t.) The best customer service agents err on the side of compassion in instances in which automation has failed. What most likely happened in this case is the organization either purchased or “borrowed” a list from some event organizer I had attended during my lst trip to D.C. and automatically added all the emails from it. In any case, moving toward a solution is better than placing blame on either side.
- Not solving the issue. I’m unsure why the organization insisted it was unable to resolve the issue even after its provider confirmed that it could, but I’m reminded that in the world of automated emails, one human being taking just a few minutes to provide service is all it really takes for a user to have a good customer experience.
Automation #2: Ordering delivery (Good)
In sharp contrast to the above experience, last weekend I had a seamless experience ordering a delivery dinner via the internet.
When I found myself too exhausted after a full day at TEDx Rainier ’11 to make or seek food, a friend suggested 24hour delivery. I’m sure others use this site daily, but I almost never order food to be delivered, so I wasn’t sure where to start! But the site made it effortless, seamless and simple, just as every customer experience should be.
First, I did a search for restaurants that delivered at that time of night near my location. Then, I checked the ratings and selected an Indian restaurant that had good vegetarian reviews. As it turned out, clicking on the restaurant name brought up an interactive menu, on which the user can click on items to instantly add them to an immediate order. Likewise, each item has the option to include special instructions, such as “extra ketchup” or “medium spicy.”
Even better, after the order is complete, there is the option to review, opt to pay in cash in person or via credit card; there is even a spot for a credit card or cash tip and for instructions on getting in to the user’s residence. In short, the automation on the site considered all possibilities, absolutely everything the user might want and the driver might need to ensure a smooth transaction.
And smooth it was! A receipt was delivered via email for recordkeeping, and the site has a humor-filled page of FAQs should something go amiss. In short, every need the customer might have was already addressed, so only the most bizarre of situations would merit contacting the company.
What went right:
- Anticipated customer needs. The automation worked because every aspect of the search, order and delivery process had been designed as if it were a science.
- Gave more information, not less. Instead of making it difficult to report issues or assuming the customers would call if they were upset, the site proactively listed the most common questions in a friendly, laid-back manner, undoubtedly reducing the number of complaints both to the retaurants and to the site.
- Provided the opportunity to give positive feedback. The site allows users to rate the restaurants and create a community feel rather than a top-down situation.
Two examples of automation: one email automation that banks on users being too lazy to notice the involuntary robotic subscription to complain and one e-commerce automation that anticipates and delivers users’ needs to complete a transaction.
Customer service automation isn’t evil; well-designed self-service experiences can be incredibly easy, stress-free and provide a better customer experience than even an interaction with a live agent. It’s in the way the organization chooses to address customer needs that makes the difference.