If you follow my blog and my writings (and rantings, and presentations, and panels — if you ever talked to me about this) you know that I am not a big fan of using Twitter for Customer Service.
It is not that it is not possible to do it well, but it is that the resolution times, close rates, escalation rates, and just about any other metric you can use are so horrible by comparison that to do it is almost a waste of time and resources.
This prompted me, about a year ago, to write a post advocating the use of a single channel strategy, and even before that to deliver a presentation on the failing metrics of social channels.
Although things have improved, somewhat, for smart organizations that have learnt along the way, my core statement remains as it was at the beginning: Twitter is no more than an appropriate triage tool for Customer Service (I think I called it an IVR back then, I still do today).
Midst the poor performance and lack of understanding from organizations though, few glimmer of hopes are emerging.
Here are two examples, in pictures, of companies that are getting the gist of using Twitter for triage and escalation when necessary – and have the right tools to do so (which is the hardest thing to do using Twitter for Customer Service BTW).
Example One: T-Mobile Escalates To Chat.
In the picture below you can see a customer asking for help with a billing issue. Now, there are two bad ways to handle this: 1) ask the customer to call and give them a ticket number (after asking them to follow you, DM back and forth), 2) try to resolve the issue via Twitter (yes, even via DM) in 140 characters at the time.
Alas, T-Mobile did it right – realizing it would take more than 140 (or 280, or 420 — yes, I did take math in college) characters to resolve it, they immediately escalate to chat.
Why this is better than calling or emailing?
The customer wants immediate resolution, more than likely, and they come to Twitter for that. By escalating to a real-time channel (chat is one) that is easier to use, less expensive (on average) than phone, and can be even outsourced without major issue (as opposed to the telephone being outsourced and customers complaining) they can control the SLAs, the privacy of the customer, and the wishes of the customer. Even if the customer wants attention, and not real time resolution, the offer is a good way to set expectations: we are here to help, in real time.
BTW, I clicked on the link, it worked – but it was time-sensitive and expired shortly after it was issued – even better.
I also imagine that the chat session would show up in the unified desktop that T-Mobile agents have, where they will get access to KB, customer history, etc. Likely better than the tools they have for Twitter (educated guess based on what I know they do).
Example Two: Amazon Escalates To Web-Based Ticketing.
In this second example a customer complained about something that was not right with a product made by one of Amazon’s companies. They quickly replied with a link to provide additional information.
The interesting part here is that (if you notice, my name is at the top of the screen) by doing this Amazon can see the type of customer I am (I am prime, and I use it very often), what products I purchased, when, and other information they need — in addition to being able to link my Twitter ID to my Amazon account (if not done before).
Social ID correlation is a huge, huge, huge problem for companies — and this is an easy solution to that problem if customers are logged in.
Bottom Line: Learn how to use each channel properly. Social channels are horrible for resolution (even if you get past the 40% of unnoticed events, the 10-20% average close rate, and the 10x or more resolution times) and they are perfect for triage and escalation.
Do it well.
What do you think? Other examples of well done Customer Service via Twitter? That is scalable? Viable? Sustainable?
Would love to hear you thoughts…
13 Replies to “Twitter for Customer Service? These Companies Get It Right”
I agree, Esteban. Social has a “real-time” feel to it, but real-time is awfully difficult to scale using traditional models and technology.
Two factors jump out at me: know your customer and manage expectations.
Amazon moved the conversation to a platform where they know more about the customer. That’s one way to handle it. Customers seem pretty eager to share information with vendors, now vendors need to show they’re listening when it comes time for customer service.
A larger issue (perhaps) is expectations. If you (as a vendor) don’t give guidelines for expectations, your customers will set them for you. And that makes it really hard to meet them.
Customer expectations – the final frontier— well, not really. But you are hitting on something pretty crucial and something that I have written plenty about before in this blog (don’t want to spam links, but glad to point it out if someone asks).
Here is the deal, when it comes to expectations it is always, always best to underpromise and overdeliver. No questions about it.
As you say, if you don’t set them – they will set them for you. In the case of twitter, companies without SLAs (expectations management contracts would be a better label) for the channel will fail miserably. Companies with SLAs (and they are the vast minority) will find ways to use as a triage vehicle, escalate as appropriate, and use other channels to complement.
I happen to know that both of these companies have SLAs for social channels. Thus, they manage expectations… now, if they would just make them public they would close the loop much easier and be ahead of the game. Step-by-step.
Thanks for the time to comment.
Nice post E – thanks for sharing your thoughts as usual – insightful.
I will simply pose a question: Do you think that (or have you seen) Companies do Customer Service Journey Mapping? You know, the type of Journey Mapping that has become so popular on the flip side…
(For example, as I pondered which social login to use to post this reply, if AMZN strongly suggested adding your twitter handle to your account info, if you like to use twitter for service…the mapping become easier, no?)
Thanks for reading and the question.
Journey mapping is quite popular these days, most people think it will help them understand the customer’s expectations and the customer’s mind. It doesn’t really, but it does make for a nice way to say “we care about the customer”.
I have written many times about this, but the problem with any concerted, organized effort to try to understand the customer (which is actually not a bad initiative) with the purpose of then creating “the perfect experience” (which is where it goes to — well, you know).
I have talked to so many people lately about Customer Journey Mapping (still mostly driven by vendors and consultants, somewhat skeptical end users have not fully embraced) and what it means in a cross-channel, omni-channel, or even multi-channel setup that I cannot say I am encouraged by the use of it.
In a single channel scenario, it sort-of kind-of can work – that is at least until the customer changes their mind on how to get what they want or need, a single channel does not meet expectations, or their expectations change based on external factors (from your competitor does it better to i don’t have time to wait today).
We can have a very long discussion about this, but the problem with mapping journeys in multi-channel environment is the lack of a common data model that allows the organization to track an user across channels and effectively deliver an omni-channel experience.
In simpler, non-hype words: it does not work in complex setups, and all customer service setups are becoming more and more complex as social channels, mobile interfaces, and reduced wait times driven by increased expectations enter the fray.
If you have a 60% escalation for social channels (which is superb and have only seen it a handful of times) you still have more than half of the interactions going to another channel – how do you keep track of that and account for it in journey mapping? especially when it is not predetermined by user or experience.
We can go on forever, but with journey mapping – good idea to start to understand the customer, but not the solution that is being proposed.
but that is fodder for another post…
Thanks for letting me ramble on…
I like that you’re calling out good examples of service through Twitter, but I wonder why you think it’s generally a waste of time. You seem pretty sure of yourself, but you also seem to imply that customer service on Twitter is made up solely of special situations that are better handled privately. Given that most service requests are not complaints or sensitive issues, I would argue that customer service in social is both quick and easy while giving the brand a touchpoint to demonstrate value to the customer.
As a sidenote, the specific chat function TMobile is using doesn’t work well (or at all, really) on mobile, so I wouldn’t recommend it until that’s solved.
I am not sure you follow my position well,
I say that social channels are excellent triage channels and should be automated.
It is a waste of time not to do it that way.
Big diff from saying it is a general waste of time.
Alas, since most organizations cannot do them properly, then better off not doing them.
Escalation is the biggest issue with social channels (that 85% you mention in your next comment, to save time and answer them both here, is average – that means that some people will have just 40-60% (best i’ve seen) and some people will go over 85% (most of them, btw)) and it is closely followed by non addressed issues (average is 40%, again – average).
You are looking at success stories and making them common place, for each of them there are 12-15+ more that are not success story. For each T-Mobile replying to all tweets within 45 minutes and resolving over 60% over original channel there are 5-10 organizations not staffing them 24-7 or not using the tools to pick up more than 50% of mentions.
Any channel you do poorly is a waste of time, that was made apparent and clear in the post I wrote for desk.com. I don’t care if you commit to doing just Twitter. Phone, Email, or something else – as long as you can master it. Only then bring in a new channel, and another, and another. Suggesting anything else is not only irresponsible, but opens the organizations to downplay the excellence they have achieved by doing it poorly elsewhere (try to get service via t-mobile.com – you will have to go back to another channel – they suggest phone usually).
Anyways, not picking on any one, just relating one that I know somewhat — have many other examples.
Bottom line: only those companies that understand single channel excellence can “easily” succeed at bringing in new channels – and even then, only for the best use for each.
Thanks for commenting,
This examples are goog presentation for the ideal customer service. But most of the problem are not general questions. I am dealing since 11 year with customer service centers and there customers. Most are specific customer contacts and special questions, which needs customer data. And no company should ask the customer for sensitive data over Twitter and Facebook.
Yes, customer service via Twitter and Facebook is possible, but if you read und follow many great companies, like Deutsche Telekom, T-Mobile and many others, they allways transfer the customer to a different channel.
So the customer notices, that he gets an answer, but after that, he must use another, company owned channel. More then 85% of all customer question via the Social Networks will be transfered to other channels. And the customer learn one thing from this, use the other channels directly.
With one execption. If you habe a complaint an need the help from the community to reach your target, use the Social Networks to get a high pressure to the company.
“More then 85% of all customer question via the Social Networks will be transfered to other channels.”
The validity of that stat is my concern. I don’t believe it’s accurate. Either that, or the support people are doing a bad job by unnecessarily transferring to private/external channels.
My experience is limited to only one company, but it’s an enterprise level organization with a major focus on customer service. In my experience, the majority of interactions remained in the social channel, but that was also because we had real support people working the social engagement. That’s how you need to treat a support channel. If you have other folks managing the social channels (folks in Marketing or PR or untrained interns), then you’re setting yourself up for failure by putting non-support people in direct contact with the customers who ask for support.
I worked with many diff size and diff industry companies over the years. That is an accurate, and not very fast moving, stat. Two years ago it was 92%, so we may have improved a tad.
could not agree with you more, actually – i have similar stats and statements in the presentation i linked above.
many thanks for confirming my thesis. the non-case study companies are the ones that bias the results for all others.
Very informative article and interesting critical view of using Twitter.
@Alon: I think the main problem, which drives companies to transfer the contact from Twitter or Facebook to other channels is in the most times the data privacy. No company will be asking a customer for his sensitive data like customer number, bank account number or anything else at a not owned data environment.
I must also mention, that my stats came from companies in germany. Up from the data security law we handle customer contacts very strict. But Esteban agree with my information. So i think we will also have the same problems in other countries.
The discussion about NSA oder GHCQ didn´t make it easier for companies to handle the whole request within the social networks.
On the other hand you must also remain, that the customer decides if he wish to use the social networks.
I have wrote last year an article about two great german companies and the customer acceptance of the social networks.
The article is only in german, but i think, if you look at the charts, you will see, that service via Facebook and Twitter is not to top support channel today.
I’ll have to accept that my experience may be an outlier. At the company where I managed social support for over five years, we were in the fairly unusual position of being able to frequently identify specific customer accounts without asking for protected personal information. That may account for my personal impression that you don’t need to transfer to alternate channels so much. I maintain, however, that there are loads of contacts made that have nothing to do with account specifics. The idea that this type of interaction accounts for only 15% of activity on corporate pages is crazy to me. Still, I don’t have hard data to support that position, so I can only say it ‘seems’ crazy… I don’t really know.
Comments are closed.