Scaling Social Media
August 14, 2009 Justin Kistner — 20 comments
Jeff Katz recently sent me a link to a post from a blogger complaining about the state of technology in social media monitoring. While I agree with many of the points in the post about the challenges of the industry, I was very put off by the insulting manner the blogger used. I was also surprised to see how many start ups in the social media monitoring space seemed to just learn the challenges of the market.
In a comment string on that post, Ari Herzog asked me to elaborate on this point: “If you think the amount of person hours spent using the current systems is a lot now, just wait for what is coming.” So, this is that elaboration.
Staffing up
Jeremiah Owyang will be the first to tell you that he doesn’t think social media scales. He’ll also point you to Best Buy as an example of a company to watch. Have you seen their Twitter background for their @Twelpforce account? It’s an army of people. Best Buy is doing something unique: they’re solving the volume of conversation with, gasp, people!
Most of the business interest in social media has been around marketing. As a social media marketer, I know it’s only a part of the long term business value of social media. The truth is social media is a communication channel, which plugs into every business department just like email and phone. Speaking of phones, former CEO of Zappos.com now acquired by Amazon.com, Tony Hsieh said, “We don’t really think of social media as a marketing channel; that would be kind of like asking about ROI on answering phones.”
Having started by building social media programs with systems that I rolled myself to building out a global social media program for Webtrends; I can say that a big realization has emerged from what Jeremiah and Tony have been talking about. Businesses need to think about social media as a communication channel. They should be planning for it and evaluating it using similar KPIs as they do their phones. It’s not a question of if they should talk to customers, prospects, partners, etc. The question is how efficiently they’re doing it.
However, that’s not what businesses are currently thinking or doing. Most businesses are still sticking their toe in the water of social media and as a result have seriously under invested in the space. Large Fortune 1000 companies are gripping about spending $30K on software and handing it over to super small teams–sometimes a single person. Can you imagine if they tried to answer their phone lines with a 3 person team? Or a single intern?!! How can they expect to evaluate performance when the team is so under water that they can’t even think?
The reality is that if businesses want to be successful in participating in social media, they’ll need to allocate resources in proportion to the volume that exists for their brand. Small companies can get away with small teams. Large corporations will need large departments. They will be structured like call centers with IVRs, scripts, answer trees, etc. It will take a substantial investment in staff training, infrastructure, and rolling out business processes. Everyone in the company will have to know how to use the tools and different departments will be responsible for different pieces of the conversation.
There is wide spread confusion about how much to invest and how to measure value. To help, Forrester or someone should do a study comparing the amount of conversation hours large brands have on phones, email, and social media. That would help businesses understand the relative investment to make. It would also help them determine which channels are most efficient so they can push conversations to their most efficient channels. We’re deploying a cross-departmental, multi-channel, global social media program here and we plan to share the details for our customers’ reference.
CATEGORIES: Industry Perspectives
TAGS: Social Media
20 Comments
2009-08-15
05:31:25
Justin,
I followed this post as a trackback from Asi's. I think Asi's post was a matter of fact account of one users experience with several monitoring platforms which made him feel considerably underwhelmed. Surprisingly, I noticed two of the four vendors named chime in and accept his call as legitimate reasons to strive towards making their product/service better. Harsh as his words might have appeared, I think we need to look at those concerns with monitoring technologies using the same rules of engagement and a willingness to take corrective action. As a vendor in this space, his opinions resonate with me particularly because when we treat such harsh opinion with disdain rather than listen and treat it as anything other than user feedback, we allow a false dichotomy to further perpetuate in this space. At its most basic level of understanding, if the “thought” is to deliver social media monitoring tools that are meant to accelerate the analysis portion, then lets focus instead on the delivery of results and objectives that can be backed-up by "action" rather than catchy metric lingo and promises on ROI that can’t be kept.
With regard to your points on scale, you piqued my interest with this idea of assigning a human touch point to brands. As it is now, when we monitor brands on behalf of clients, the indicators for outreach and engagement depend on some metrics like sentiment, geography, source type (ie. blogs, forums, etc). While each clients requirements and strategies are different, comments and context are still the proxies for outreach and engagement. But if we used sentiment scoring as an indicator to make a crude example, lets say Brand XYZ decided it only wanted to engage with negative incidents. Here is where I see your points on scale causing disruption to Brand XYZ's social media engagement strategy in the future. What if in 5 years Skype becomes a contact standard (just noticed it on Kijiji listings) where each social media incident (positive, negative or neutral) could potentially be signed with a "Call Me" request? Some might see this as nirvana in online brand building and the forming of new social media relationships because each incident has an actual person on the other end and can be handled in a timely, dilligent and offline manner.
The reality of an altenate future like this would mean a decentralized consumer inquiry channel allowing consumers to ask questions on pricing right through to real calls of concern - and may I say with an even more evolved expectation on earning their trust and satisfaction. In other words, you could forget about responding to a negative incident a day late. In such a case, I don't think it would be relevant to know Brand XYZ's size to properly scale for solutions. The question is whether even brands the size of Best Buy, who effectively use the imagery of "people as team" for a Twitter background, would be just as effective in delivering on the manpower and human resources needed to deal with incidents in that kind of alternate future? Perhaps by then we might also have an evolved reputation management situation where the social Web is inundated with so many human transactions and experiences that the negative incident ratio of a company or brand becomes less important in informing consumer choices.
Thank-you for advancing this follow-up discussion!
Joseph
@RepuTrack
2009-08-15
13:13:30
Joseph, I think it was a matter of fact account too. And, you're right that it should have been taken as customer feedback, which is nothing to have disdain for. Not only is raw feedback valuable for companies, but more importantly we have to be able to talk about life from the perspective we are experiencing it. But, once numerous vendors and high-profile social media practitioners commented, it became more than just a post with customer feedback, it became a conversation with numerous stakeholders. Many people have seen that post and many more will see it still. If what they see are brands apologizing to a guy who said they all suck and that their technologies are stupid, that's not demonstrating good customer service. Do we want to send a message to all those that read his post that it's ok to be in the middle of learning how to use two services, get demos on a few more then lump them all together and call them stupid when you feel stressed from being behind on providing services you don't fully understand yet? I don't.
Having said that, it bears clarifying that Asi didn't do anything inherently wrong by venting his frustrations, even if he was rude about it. It was the pandering that was the problem. Someone needed to step in and set the record straight. I'm of the opinion that vendors saying "Thank you sir. Can I have another?" condoned bad behavior. I'm confident that he isn't pondering the words of those vendors who apologized. My bet is that was people like me who called him on it that lead to him apologizing. That was mature of Asi and clears him of all transgressions in my mind. We all make mistake and get caught up in moments, myself definitely included :) And, we all are better off from being able to read the conversation that flowed from him reporting on his experience because it is, unfortunately, an all too common one. So it was good that he posted what he did. As @kire said, "@justinkistner well something is obviously working if this fellow can garner so much attention."
To continue forwarding the discussion, I wanted to touch more on sentiment. I subscribe to the belief that viewing sentiment as positive, negative, or neutral will get people into trouble. That false notion is spread by vendors who want to sell people on the idea that an algorithm can automagicly analyze all of the conversations about their brand and categorize them into three easy buckets. Buying into that lie is a big part of what fueled Asi's frustration when he discovered it wasn't true. It's that same disillusionment on sentiment that lead the vendors who push sentiment to incorrectly apologize to him because his post registered as negative on their systems. Enabling a frustrated customer to continue to be confused because we can't look at his experience any deeper than "negative" is the real disservice here. Asi didn't need an apology, he needed help. Sometimes that help is correction.
Brands have a responsibility to their customers as a source of knowledge and experience for them. That is especially true in social media. There is a lot of noise out there and people need direction from knowledgeable people. Technology can help make conversing with the market more efficient, but at the end of the day it takes a person to see the real story of what happened with Asi. It takes a real person to respond to him as well. And, it takes a lot more people to talk with the rest of the market with the same level of nuanced care.
2009-08-15
19:24:15
Let me take a pass at boiling this down for ya'll:
Whatever vendor Asi is using oversold him on the capabilities of the system, and he is venting about getting snowed, and he doesn't even know that that's what he's doing. It's blatantly clear from his follow up comment:
http://no-mans-blog.com/2009/08/05/the-problems-with-social-media-monitoring-technologies/#comment-240532
Further to that point, based on the original post, I'd be willing to wager a significant amount of cash that Asi is not actually the one reading the actual words people are writing about his client and thinking about how to act on them. How implications touch various parts of the business, etc.
In all liklihood he is just looking at the report the vendor spits out and trying to shape it into something meaningful. A failure typical of people that come from media planning:
http://www.linkedin.com/pub/asi-sharabi/2/764/187
He should read here about the deeper implications of the connected consumer and what they have to say in public about brands:
http://www.havasmedialab.com/?p=27
In sum, it's intensely ironic that he's making the complaints he's making. Of course, having a PhD in social psychology, he's a smart guy and will eventually get hip to all this ;-)
Any questions?
2009-08-16
14:31:34
Justin
Thanks very much for taking this debate further
I cannot agree with you more yet can argue with you forever so let me try to recap some of my points.
Of course SM monitoring should have released their beta version but they should have also been more honest about their capabilities and quality of service. You could almost argue that their marketing messages are misleading and my post was indeed a classic consumer's rant: semi-structured and leaning on the emotional due to frustration from the service rather than a full comprehensive rational expert's review.
As i stated in my comment to the unexpected stir my post has caused, it's the promise of self-service analysis tool that is very problematic. The reality of the technology is that it is simply an aggregator and with some brands/keywords in my experience, not such a good one. I'm not expecting these tools to replace my analysis but I didn't expect also the amount of hours i had to spend cleaning and correcting the data from spam and irrelevant items. Specifically, I found sentiment analysis and regional filtering to be so unreliable that it practically prevented me from using it quantitatively.
Unfortunately these limitations are not presented to consumers of these tools nor is an explanation of the historical context in which these tools operate.
To your point on scale I cannot agree with you more. I usually tell my clients when it comes to the social web that they must not do it half-heartedly. "Do the Dell or best buy or don't do it at all" I want to say but that would be plain naive. The reality of human and business behaviours is that change is slow, takes a lot of time and brands naturally tip-toe before committing to real change (and investment). Perhaps the truth is that unless you get burnt like Dell (hell) you don't really get the importance of the social web as a channel for dialouge.
Finally your observation that my post was a classic example of negative sentiment is absolutely spot on. I did not elaborate on how shallow and superficial these vendors training of their clients- none of them has done enough to help me use their services in the best possible way. Half hour webinar and few video tutorials are an industry standard and the result is inevitably frustration.
Thanks again for your insightful comments
Truly yours,
Fringe blogger ;-)
2009-08-17
10:17:38
This has to take the cake as the most meta conversation ever
2009-08-17
10:25:20
Doh! Fringe blogger was a little emotional, wasn't it? Sorry about that :) Glad to see your comments here, Asi.
Spot on here. I think this is what Ethan means when he says, "Whatever vendor Asi is using oversold him on the capabilities of the system, and he is venting about getting snowed..."
I do feel like this is where the social media monitoring industry has let you down. As you said, "Unfortunately these limitations are not presented to consumers of these tools nor is an explanation of the historical context in which these tools operate." Not only are we failing you there, but also in training, which you point out, "I did not elaborate on how shallow and superficial these vendors training of their clients- none of them has done enough to help me use their services in the best possible way. Half hour webinar and few video tutorials are an industry standard and the result is inevitably frustration."
If you would have had the context and training, I don't think you would have been frustrated. I'm going to do a follow up post to this because I think we have an opportunity here. Your experience echos many others in your same position. They are struggling with detailed tech capabilities and because of something else you pointed out, "The reality of human and business behaviours is that change is slow, takes a lot of time and brands naturally tip-toe before committing to real change (and investment). Perhaps the truth is that unless you get burnt like Dell (hell) you don’t really get the importance of the social web as a channel for dialouge."
That slow moving human/business behaviors are exactly why companies are under investing in the space (if only I had a dollar for every time someone said, "I guess I'm just old school"). Under supporting social media is why people are having bad experiences, which is leading people to think it's a technology failure. Watching social media programs fail due to lack of resources is frustrating me. So, rather than all us living frustrated, let's use this moment to help ourselves and others.
I'm going to do a post on the strengths and limitations of Webtrends Social Measurement, which is powered by Radian6. I'll see if I can get some of the Radian6 folks to participate with me. I'll then provide some reference for what kind of resources should be allocated to use the technology correctly, which will include time, money, and staffing. We can use that post to encourage others to do the same. Benchmarks will help businesses understand how many resources are required for success and can decide how to invest accordingly.
We also have a training program coming, but I don't think I'm allowed to talk about it much yet...shhhh!
2009-08-17
11:15:25
Hey Justin,
I'm looking forward to your further exploring of training companies to use the tools Webtrends sells them. Interesting stuff.
The idea you brought up about social media call center style teams doesn't seem to sit well with me. Call centers are traditionally staffed by lightly trained employees without the ability to make important decisions. I think that you can probably do more with the hub and spoke model of social media engagement. A small team that filters & analyzes customer reactions (real ones, not the sentiment measurement - blech), then sends those reactions to be dealt with by the appropriate department.
The spoke and wheel seems to scale better and get a company working together internally rather than continuing the silo effect.
2009-08-17
12:32:19
@cory
the whole point of silos is scalable efficiency: specialization, etc. but to your point, hub and spoke is definitely compelling from a institutional design perspective. and it matches with type of info out in SM. but my (and i think justin's to some degree) point is that a "small team" can't deal with the sheer volume of information being created in social media.
there is a huge gap between the [valuable] information being created about a given company/product and that company's ability to ingest, make sense, and act on it.
justin has written in the past about "context engines" and it's an interesting idea re: how to automate some of these processes.
just some thoughts. great convo here.
2009-08-17
12:50:19
@ethan I see your point. Perhaps it depends on the type of organization. Zappos has their 400 employees doing their social media work & they use it as a CSR & communication channel. Starbucks, on the other hand, has a six person team linked to the rest of the company via the spoke & wheel pattern.
That small team couldn't, and shouldn't, deal with all of the UGC that SM develops. They're a hand-off team that points to the client and says, "Hey, you guys should pay attention to this over here. Go resolve, promote, or address that user."
That said, I really LOVE the Zappos & BestBuy approaches. I really look forward to seeing how BestBuy's @twelpforce plays out.
2009-08-17
13:25:13
When I was thinking call centers, I was thinking more about structure than location inside or outside of the firewall. That said, it's a very interesting discussion to look at methodologies for handling high volume. Comparing hub and spoke to outsourced social media centers would be a great study!
2009-08-17
14:36:16
Ya frankly I'm dying to see a call center that adds SM response to the responsibilities. Would be teh interesting
2009-08-17
14:58:24
Call centers could implement social media response - but you'd have to have a well trained group of people. Would it be worth the expense that it would take to properly train? Depends on your hiring & training practices, I suppose.
2009-08-17
15:05:11
Great to see this active commentary - some great insights and opinions.
Justin, I wanted to clarify something regarding your statement on sentiment - namely this:
I agree on the automagically part, but on the getting people in trouble part - well, it depends. If you rely on a machine analysis to interpret tonality, and assess risk or influence, that's where you and I can agree on the disillusionment part (what I referred to as the false dichotomy). And I think we can also agree that this might partly have fueled Asi's rant. However, it would be a disservice to firms like ours to simply lay down and accept that sentiment isn't important, or to not address the blanket statement that sentiment analysis isn't being done properly.
To read more on my take with regard to entrusting machine's with the assignment of interpreting influence and risk, please read this old blog post. In brief, and for clarification purposes, our firm does pre-screen each incident of brand mention using machine analysis, however the finalized scoring is all performed through human review, which is included in all our tiers of service. Criteria used for assigning sentiment is based on the subject of the post, the content/message, the handling (ie. follow-up comments). In addition, we have a metric category called the Eco-Graph which essentially acts as a risk radar for brands who find themselves at the center of environmental topics which are impressing or angering audiences. Incidents falling into the Eco-Graph category are scored on tonality as well.
The human review translates into a more qualitative assessment of risk, because we watch the videos and read every incident. To further illustrate how our approach to sentiment scoring differs, I will use an actual incident without naming the brand. A blogger posted their experiences with a big box retailer - specifically detailing how they were thrilled with the delivery/servicemens diligence and proficiency. The incident was carrying the momentum of a "positive" customer experience, until they remarked at the end of the post about their surprise with finding out the delivery men had left all the cardboard and foam inserts on the front lawn.
What was even more surprising is how comments began harping on the environmental irresponsibility of this retailer. In conclusion, the incident was scored negative on sentiment, and on our Eco-Graph. Keep in mind that the comment tracking was a decisive part in determining the negative direction the incident eventually took. For brands who strive to be good, responsible corporate citizens, awareness of any incident questioning stewardship would be important. As the maker of all our monitoring tools, I can't say I would trust any algorithm or machine analysis to handle the assignment of sentiment scoring, and unless its reviewed by a person, then as you put it, that's when it gets people in trouble.
Joseph
@RepuTrack
2009-08-17
15:52:40
Cory, I'm thinking social media call centers would have trained staff armed with scripts. Call centers use these FAQ style databases to help them search for answers while working with callers. They keep notes on the knowledge base and edit on the fly. There are tiers of folks to handle different levels of questions. It is a big investment, but it's cheaper than having everyone in the company answer the questions while trying to do the rest of their jobs.
Joseph, there are two issues at play here. One is about machine vs. human scoring. What you said about human and machine processing is great stuff. The combo of machine, then human will be one of the primary ways to achieve greater efficiency in scaling conversation that will translate to competitive advantage.
The other point, and more important one, is about using positive, negative, and neutral as the categories. It was that categorization that I wanted to warn Asi against. Those three labels are just too limiting to provide good analysis of social media. I would be interested in a more complex description of sentiment that covered the full range of emotions used in our languages; such as love, like, hate, protest, etc. Otherwise we get a narrowed view of what's going on that can mislead our decision making.
2009-08-17
19:20:03
Justin,
This is such an insightful post with so much to chew on. And you've touched on something that I believe so fervently; scaling your human outreach to the volume of your brand dialogue is the only true "scaling" that social media will ever have.
Mechanization - the tools - are only the bridge building. They can help humans get to certain points faster or more efficiently, but the very essence of social media is the human element, and the only thing that can absorb more and deeper relationships is more PEOPLE.
My hope is that while the need for people increases with social media, it's efficacy and cross-disciplinary nature (touching customer service, sales, marketing, PR and loads of other things) will free us up to jettison old, outdated processes, procedures and approaches so that it's not so much ADDING to what we're already doing, but evolving, changing, and discarding the obsolete to make room for a different way of doing things.
Thanks so much for your articulate, level headed discussion. I've got a whole lot of respect for the way you've handled this entire discussion, and I'm looking forward to continuing it with you. We're open to collaboration anytime.
Cheers,
Amber Naslund
Director of Community, Radian6
@ambercadabra
2009-08-17
21:13:34
Thank you, Amber. When I was saying I'll see if I can work with someone at Radian6 on a follow up post, I was thinking of you. So that's cool that you picked up on that. Very human of you :)
It would be cool to connect and discuss a follow up post to this one. I like what you said about social media being an evolution of what we're doing, not adding to the current mix. Looking at details of how that looks in terms of daily activities, org chart, systems, processes, etc. This could get to be a crazy huge post that may need to be a series, but we can talk through all of those messy details.
2009-08-18
01:12:28
[...] think of social media as a marketing channel”: Scaling Social Media « Webtrends via [...]
2009-08-18
12:49:45
Social Media Analytics is a harder sell in most organizations because the case study work justifying spend in the social media channel are few and far between.
As the social media channel becomes mature, dollars to measure that channel will flow.
However, companies that create great value where none existed before are the true innovators, and there's great first mover advantages to be had. Best Buy is a good example of innovation in action, and they're to be applauded.
Others will say 'me too' going forward.
2009-08-18
13:32:10
Justin & Amber - I look forward to the follow up post.
I work at a company that employs a lot of call center employees. I'm intrigued about the idea and am thinking of ways to implement. (gears slowly grinding...)
2009-10-07
12:28:08
[...] — Scaling social media [...]