I am in the process of evaluating a horde of monitoring tools and applications as the time has come to move on from the freebie model we kludged together two years ago to something that will support the organization’s global interest in Social Media and reputation management.
So, here is chapter two in my Social Media Marketing 201 series – how to monitor buzz about your organization for no money down, and how to know when it’s time to hit up the powers that be for some real money to spend on a real solution.
Free: yup. It costs nothing to detect the chatter about your company. There are two solid solutions for blog search – perhaps you’ve heard of Google? Thought so. Google Blog search is a good thing. And then there is Technorati, which sort of defined the space. Both are great tools, but you can automate searches of specific keywords and phrases and then syndicate those searches as RSS feeds out into a blog reader such as Bloglines or (in my case) Google Reader. Then you just need to remember to scan the blog reader a couple times every day.
Not to drop a coprolite into any vendor’s soup, but beware of claims that they detect stuff better than Technorati or Google. They are basically selling an aggregator/reader and offer some stuff above and beyond like influence maps and sentiment metrics in a shared space so you and your team can look at the same stuff. That’s great, but maps are eye candy (sorry map people, I really don’t need to see some spiderweb of connections) and sentiment metrics remain, for the most part, mainly binary sucks/rocks, green/red statements that do doodly squat to drive an action.
So, go free, use Google Reader, syndicate in some blog searches, manually sub some feeds from communities and forums, and get ready to weed through a ton of noise to find the signals that matter. And, one last thing: hiring an outsider to do your monitoring for you is suicide. You positively absolutely should monitor stuff yourself. Be paranoid, get in the habit of fretting about the stuff you are missing. It’s a good place to be.
Paid: When do you need to pay for a social media monitoring tool? Let’s assume you are tracking and issuing a report to somebody on a weekly or daily basis (note to self, blog in future on reporting formats) on what you’re seeing. Sooner or later, if you do this right and if your brand and industry attracts a lot of SMM activity, you are going to find the distribution list expanding. More people are going to want to know what you’re seeing. And pretty soon they are going to want to directly monitor the buzz themselves. The primary thing a free/homegrown solution won’t do is permit a workflow model that will let the person doing the monitoring open an incident and assign it to someone best qualified in the business to resolve or answer it. Don’t underestimate the breadth of the people who will sooner or later need to know.
Workflow – the system for easily distributing issues is one area a paid “enterprise reputation management” system should excel at. Enterprise systems are for the big guys – not necessarily the customers with high volumes of detected hits, but ones that have a complex organization scattered over multiple timezones and continents. It’s one thing to see it, it’s another to start a timer, assign, and then resolve it. Workflow is the big need we have right now.
Response: the mechanism for responding and working with a forum poster or blogger is sometimes built within the tool. Nice to have perhaps. I have never tried such an integrated system, but I would be hard pressed to call it a need-to-have.
Metrics – as I have said over and over, nothing in this world is more Dilbert than metrics for the sake of measurement. Sentiment metrics are particularly vexing, volume sentiments are crap. Be wary of tools that purport to deliver metrics. I subscribe to the manual scoring method of sentiment. If it really matters then the only viable way to tell whether a post is positive or negative is to read it and make the call. Doing so will entail enlisting the support of a team of monitors or moderators, or, if you are a sole operator, doing it yourself.
Nathan Gilliatt has compiled an excellent Guide to Social Media Analysis
and could teach the PhD level series on this topic. I’ll report back as we go into deeper trial with the various solutions in market and let you know where we end up. Thankfully I have a great team in Mark Hopkins, Tim Supples, and Esteban Panzeri to guide us through the thickets. Their statement of requirements is very solid and if they agree, I’ll post them in a subsequent post.
Â
Â
Â
Are you a Forrester client? If so, I can provide a wealth of in-depth information on the selection process.
Peter, I am a Forrester client but don’t have a seat per se. Can we discuss offline?
Wow, David, thanks! I wouldn’t presume to teach PhD courses, though. Maybe the executive education program at the B-school. 🙂
sorry for v.brief response (from mobile)…but this is excellent. would be great to see the 301. using workflow to attack the seams between bizunits. once in place what would an ops scorecard look like. mining for and engaging influencers. integration w/ existing systems and process.
great post!
sean
I’m still in the realm of free monitoring myself. It does work but, its definately getting more difficult. I’ve started looking at various vendors, I’m sure the same you’re considering.
I echo Sean’s comments and welcome the 301 course myself. Using workflow to “attack the seams between bizunits” is a must have if one wants to really get at the value of what’s happenng online.
Is it possibly to connect what’s hapening both in front of and behind the firewall?
Mike,
It seems to me that one of the challenges with Free is that by varying keywords and doing some recursive searches you can find a lot more. In fact, you can find so much stuff that it is downright impossible to quantify into metrics that one feels comfortable saying are really indicative of the reality out there.
Free does seem to find the stuff that pops up in other people’s alerts. That in itself is another challenge – the impending contest for determining what counts and what doesn’t. Say, you get a spiffy new tool and begin to generate reports, and someone else does a random google searh and finds something you didn’t. One can get sucked into a debate about what is the the true nature of the blogosphere at any moment – how many posts are there really about you? Is the one someone else found counted in the results you have? Should it be?
Even if you go paid service, perhaps keeping some google alerts running may be beneficial in understanding what your exec team is seeing and will be reacting to (since many execs now have discovered google alerts themselves).