The Burning Irony of Salesforce’s #NA14 Social Media Nightmare
“Okay, let me tell you the difference between Facebook and everyone else, we don’t crash EVER! If those servers are down for even a day, our entire reputation is irreversibly destroyed!”
— Mark Zuckerberg in The Social Network
Yesterday, Salesforce had a bad day. Three out of Salesforce’s forty-five North America instances had a power disruption and one, known as NA14, went dead for over 19 hours sending thousands of Salesforce users spiraling into a black hole of angst, terror, rage and despair.
Salesforce owns an interesting niche in the world of technology. Over 150,000 enterprises of all sizes — predominantly medium-to-large businesses — rely on the CRM software-as-a-service giant for managing the sales, communications, marketing and customer service operations that serves as the linchpin to their day to day business activities.
When a Salesforce instance like NA14 goes dead, the work of thousands of businesses (~2.2% of Salesforce’s North American clients) and their likely tens of thousands of employees come to a screeching, flaming halt.
Examining how Salesforce responds to a crisis situation of its own is especially interesting because Salesforce sells so many social marketing tools and markets itself as helping other companies become Customer Centric organizations able to receive and respond to even the most minor of service issues with exceptional and personalized care.
While many businesses cannot afford the entire line of Salesforce Marketing Clouds, Modules or Components, Salesforce itself should be the pinnacle of a customer centric organization because they at least get access to all of their shiny toys at no cost.
Unfortunately, yesterday’s service interruption, and Salesforce’s response, brings to mind another hilarious-if-didn’t-happen-to-you disaster: when JetBlue left an airliner full of 123 passengers trapped on the tarmac for over 7 hours, back on 2011. Just as those folks were begging to get off the plane, Salesforce users were begging for updates and a resolution. Salesforce users were dealing with lost time and lost revenue, living in a frustrating and maddening purgatory with no real explanation or timeline for a solution.
As per usual, in the midst of a crisis, Salesforce users took to Twitter to get answers to their burning questions; How long til it’s fixed? When can we get back to work? How much revenue exactly can we expect to lose?
Users got one of four very typical crisis communications answers:
1. The “Gee, We’re Really Sorry, Take a Number” Response
Their social media team, using Salesforce-owned Radian6, responded with form tweets (plus personalization with the user’s name). But this is the off the shelf, meaningless reply given by nearly every single technology company that deals with an outage of any kind. It’s easy, it’s fast, and it serves an important purpose — it gives an individual person an acknowledgement that they exist and matter to some degree.
On some basic psychological level, responses like this probably helps a customer to not hate the technology company who has ruined their day outright.
The take-a-number approach has a weakness though, it requires some kind of follow up.
Just as a Flight Attendant or Pilot can only say “Please bear with us, we’re trying to find a solution, thanks for your patience” so many times before the passengers start to lose their temper, Salesforce users needed to get more than placating acknowledgement.
Once the quick acknowledgement has been given, there needs to be some further broadcast message or series of messages from the company and further explanations that can help mollify the angry users. That’s particularly true for a user who might be especially angry if they got a take-a-number answer but then sees someone else got…
2. The “Hey You’re Really Important, So Here’s an Apology From The CEO” Response
How would the coach class passengers feel if they got a smile and a nod from a Flight Attendant while they could see the first-class passengers getting personal messages from the CEO?
Going through the mind of every Salesforce customer who didn’t get a personalized apology from CEO Marc Benioff is something along the lines of…well, it’s probably not printable here and involves lots of letter Fs.
At the same time, as we the audience watches the drama unfold, going through the mind of pretty much every experienced social media operator is something along the lines of …
· Why is the CEO publicly apologizing and thanking these people?
· Why is he only apologizing to a dozen or so?
· Who are these people?
· Did he just signal to the entire world who some of Salesforce’s top-paying customers are? Or are these his golfing buddies?
· Who gave this man access to Twitter?”
…especially when the follow-up conversations are also public, and include an “email me for all the details” signoff.
This kind of answer is not particularly helpful when given in public. It only serves to anger rank and file customers and explains nothing to anyone other than the fact that some people are apparently very, very special to Salesforce and that they will have access to inside information.
Still, at least someone is letting the world know that they care. Unlike this common answer:
3. The “Get All Current Details On This Outage At Our System Status Page” Response
For those who go to http://trust.salesforce.com/ and click through the NA14 service outage links, they get a fun little message, which (as of 11 PM East Coast time yesterday) read in part:
Wait…what? You’re going to recover from a prior backup? How prior? How much data are we going to lose, here? Is there an ETA on this? Hello? Anyone? Should I keep drinking, or should I just grab a toaster and head for the bath?
There are so many issues with this.
To be fair, crisis communications is a tough game. The people who write these updates are either engineers or spend enough time hanging around with engineers that they might as well be engineers. Just like the airline pilot and his flight crew dealing with a difficult situation, they’re going to share the salient facts — no more, no less.
That’s their job, it’s usually the way they think, and it’s what they’re trained to do.
We also have to understand that the communicators involved here are probably not hardcore, trained crisis operators. Although, one wonders where the crisis folks are… Have they been called in yet?
It’s long past time.
Status pages like this are generally built by technical folks for technical folks — which is fine on a day-to-day basis but in a major downtime situation can be extremely frustrating for regular people. The time stamp on the updates is UTC, also known as Coordinated Universal Time (what some people might remember being called “Greenwich Mean Time”). What does that translate to in a customer’s time zone? There’s no clock, timer, or frame of reference on the page.
These are the little things that communicators used to dealing with laypeople understand are critically important; it’s bad enough that work has stopped and data may be lost, but then to have to Google time zone conversions in order to figure out when the latest update came out adds more frustration to an already hellish day.
Dry, just-the-facts language and user experience designed for insiders are a killer in a crisis communications situation that affects regular people, but this is very typical; it’s what the front-line communicators, especially in a technical environment, are used to doing on a daily basis.
Further, the folks on the front lines tread a very fine line between saying enough to keep customers from breaking out the pitchforks and torches and saying so much that they put the company’s reputation at risk. Or worse, something that puts the company in a liability situation.
That’s why it’s generally hilarious (in a gallows humor sort of way) when the next response, though rarer than the first three, rears its ugly head:
4. The “Wait A Minute, Who Actually Released These Details?” Response
@chris_f59’s tweet says it all — “So this is a planned lack of transparency…”
When a situation of this scale and scope starts to drag on, communication strategies — if there are any — start to get fragmented. The left hand doesn’t know what the right hand is doing, and neither hand has any idea that the feet have started typing as well, and all four extremities are furiously doing their best to control the damage without the benefit of the official talking points memo (because that’s still tied up with legal, of course).
Yes, First Class passengers paid more; but in a crisis would an airline ever supply only them with warning of or instructions for a crash landing?
Is there any airline in the world that would allow First Class passengers to take on the role of spokesperson and provide the only meaningful updates ever to be received by Economy Class passengers?
Why is this situation so rich with irony it might as well be foie gras?
Because Salesforce has positioned itself as a vocal industry leader in both Social Media Marketing and Customer Service Operations.
Salesforce’s Marketing Cloud product site touts a suite of tools that will help users create “exceptional social experiences” and “drive the social conversation.”
Since 2011, Salesforce has gone on a buying spree, acquiring social media platforms Exact Target, Buddy Media and Radian 6. Each of these platforms were social media industry leaders in their own right, and when Salesforce absorbed them, CEO Marc Benioff and his team were sure they were on the path to social media marketing dominance.
Their machine was working just fine during the crisis, with a myriad of marketing-focused tweets being broadcast, interspersed with the damage control tweets:
As any seasoned communicator can tell you though — having awesome platforms, programmatic, scheduled delivery queues, automated responses to customer inquiries that look like they were typed by a real person, and laser-focused targeting is no substitute for knowing what to say and when.
This is one clear aspect of Salesforce’s ironic failure in this case.
The other issue is the fact that Salesforce has, for years, been one of the top choices for companies looking for customer service platforms. Salesforce’s ServiceCloud is positioned to do four things — close cases faster, provide smarter self-service, personalize customer care, and deliver support anywhere.
Where’s the customer support in this case?
Well, if you’re special, you can email Mr. Benioff directly for the inside scoop. If you’re like the rest of Salesforce’s 150,000+ customers in this situation, you can get the meaningless, not-at-all-helpful engineer-speak explanation and confusing user experience from the “smarter self-service” status webpage, Trust.Salesforce.com.
So what should they have done?
While it’s easy for us to satirize, torment, and even mock these poor people from the sidelines, let’s be clear: as communicators, marketers and advertisers with long experience in some of the most controversial and disliked industry verticals there are, we do all of this out of love and sympathy.
We’ve been there. We feel their pain.
We know that everyone at Salesforce — engineers, communicators, customer service reps, and of course Mr. Benioff — genuinely care about their customers and are probably sick to their stomachs that this has even happened. Pray that it doesn’t happen again; but heaven forbid, perhaps this strategy would serve them better:
1. Yes, give people the seemingly-meaningless acknowledgement and thank them for their patience.
This is the easiest kind of scripted message to put in the crisis playbook as any junior operator can push this message out with high frequency. As noted, it doesn’t stand alone. It requires a broadcast follow-up, but serves as a great first step because it demonstrates that Salesforce understands the real angst, terror and dismay being felt by real customers.
2. Translate engineering-speak to regular-person-speak and address regular-person-concerns.
Many of Salesforces customers may be able to understand the engineering speak, but just as many — if not more — will not. Further, the “just the facts” approach generally does not mollify or allay the fears of the customer. They simply don’t care what broke, or why, or how.
Just like the JetBlue passengers only really cared about when they were going to get off the airplane, all the Salesforce customers care about is what the situation means for their business.
Is data lost? If so, how much? How long is this going to take? What will the aftermath of this look like? And, what should we, the users, be doing to prepare? Will we all have to re-do the last few days worth of work?
When communicating this kind of information, remember that all customers matter. If elite, VIP clients get access to an hourly conference call of updates, then Salesforce should provide translated summaries of those calls 10 minutes later to the general public channels.
Salesforce needs to make it clear that they understand the real-world effects of this situation. Businesses large and small are losing real dollars — and not small amounts of them, either — with every minute of this outage. Deals aren’t being closed, products aren’t being shipped and customers aren’t being served. For businesses on a tight cash flow, a service outage of indeterminate length could be a disaster of epic proportions.
The key challenge here is that typically, the engineers and techs just don’t know how long it’s going to take, what’s going to be lost or saved, or what the end result will look like. It’s okay to admit that you don’t know — you just have to do it carefully. If you must resort to engineer-speak or choose not to admit that you don’t know, at least add some real-person ‘we understand your situation’ type language to those messages. And, if you serve customers who aren’t all seasoned IT professionals, make the environment in which you share your official communications user-friendly, welcoming, and easy to understand and navigate.
3. Leaders like Mr. Benioff need to be out-front and public, but to customers at large, not individually.
Updates and announcements that come out from brand channels should come out simultaneously or very closely timed from leader’s channels as well. If Mr. Benioff tweets a status update (or series of updates) once every hour or two — it shows the customers that this is being taken seriously. Customers are feeling real emotions; Mr. Benioff and other leaders will earn respect (and leeway!) by being there during those customers’ dark times.
4. Hire fast lawyers and get the pros in the game fast.
It’s hard to tell when a situation like this is going to be an easy fix or a larger problem. As soon as there’s a reasonable hint that it’s going to be a big deal, you need to sound the alarm and get busy.
Get your communications pros in the game early. And, if you don’t have planned set of talking points and an adaptable crisis playbook with a timeline, start crafting your real-person-speak talking points and responses as fast as possible and distribute them to everyone whose hands are likely to be on a public-facing keyboard. Have a consistent, user-focused message and get it out quickly.
To that end, if the lawyers need to review the talking points, and lets face it, they always will have to, make sure the lawyers doing the reviewing have an understanding of the users’ perspective, not just a focus on corporate risk mitigation. They need to be hip to the idea of being fast to market with helpful, meaningful, regular-person language. Yes, the company must be protected, and that is the lawyers’ job, after all, but the lawyers must be cognizant of the fine line between legal safety and long-term corporate reputation and repeat-business safety.
5. Have a plan and practice it.
This is so Communications 101. And, you can believe us, it’s the most annoying thing a professional communicator can hear — so it pains us to say it.
Yes, we KNOW you have to have a plan, and we KNOW we have to practice — but there’s this thing called the real world. What our professors and countless consultants have drilled into our heads isn’t always feasible, easy, or even possible in the real world.
But every business operator also knows these things happen. Servers will crash. Employees will make mistakes. Critical components can and will fail. Acts of God do occur.
And if it hasn’t happened yet, it will, sooner or later.
Having a crisis drill every once in a while isn’t a bad idea, and if it means that a number of key players are taken away from “real” work once or twice a year to practice for better crisis response… Well, it’s a lot cheaper than looking this terrible.
Again, we say these things with love and sympathy. We’ve been there, Salesforce; we know what it’s like to be on the receiving end of every negative emotion there is.
And, even though one of our marquee clients is a Salesforce customer and we’ve been having a fire drill of our own here since NA14 went down — we know that you care, and we know you’re working on fixing it but we would prefer to not have the best information coming from @timomio instead of from Salesforce.
Those JetBlue passengers did eventually get off the airplane — but none of them were particularly thrilled with the way JetBlue handled the situation.
NA14 was brought back online after the nearly 20 hour outage with a loss of less than half a day of data. Salesforce has an opportunity to continue communicating about this issue, to try to placate the thousands of customers who are facing significant loss of time, energy and money to the downtime and the day or days required to recover.
The unfortunate part of this is that, given the poor in-the-moment responses, even if they “stick the landing,” so to speak, the damage to Salesforce’s brand and reputation may already be done.
Jay Haverty is CEO and Seth Cargiuolo is Director of Communication Strategy at d50 Media, a full-service direct response marketing and advertising firm. d50 Media specializes in customer acquisition and advertising for clients in personal injury law, structured settlements, personal taxes, financial services, higher education, casino gaming, and non-profits.