My friend Marshall Sponder alerted me to a comparison created by KeenKong around conversational metrics between our two accounts when he blogged about it. What's interesting is that if you take a look at the screen shots he grabbed from the comparison for April 7 and 8, you may jump to the conclusion that I'm a total jerk - at least on those two days.
I mean, look at those numbers, I'm not out there using @ reply to its fullest or something. Well, in my view, they don't tell you much about my ability to create connections -- after all, I am the Conversation Agent! Let's dig in a little.
According to KeenKong -- what's in a name? this should be interesting to hear -- I don't respond to tweets mentioning my posts, I don't even keep using @ reply with the people who respond to me when I collaborate with other professionals on Twitter, and so on.
Could we extrapolate a trend for my account with two days? Doubtful. Yet, let's indulge in responding to the data.
If you're a business owner and already skeptical about spending time on Twitter -- actually, you probably see it as wasting time on Twitter -- you may see this kind of report as a further proof that social networks are a time sink. Given that I am on Twitter already what it feels like a lot, compared to other networks, these numbers/results would give you pause.To further complicate matters, imagine if you're reading this from an organization that is already thinking about how your employees should help you establish a presence on social networks -- and do so without frittering away time just being chatty. In the case of April 7 and 8, clearly, I was working and not chatting away on Twitter.
The CMO at KeenKong caught my tweets in response to Marshall's post and tweet. Guess what they did? Instead of reaching out to me within the network where the conversation was happening they sent me emails. Interesting choice, given the public conversation. Eventually there were also tweets. To me this means that for meaningful conversations we still choose off line communication -- email.
You should never start from the tool when you think about your marketing strategy using emerging technologies.
In a post not long ago, I shared with you the Twitter @ConversationAge effect. That is one part of how my integrated marketing communications implementation shows up on Twitter. There is much more. For the moment, if you've met me, heard me speak, or worked with me, think about how that experience was. Then ask yourself: was that experience consistent with how I interact in social media?This is important for business owners -- you're in the driver seat. Please don't take these examples and case studies as a boilerplate for your own participation and use of the tools. That would indeed lead to a disappointment for you and a lack of correlation to the bottom line for your business.
What do those numbers mean within the context of my strategy?
One more essential part of your strategy needs to be a consideration of the feedback from your customers and relational ecosystem. Over time, I received the following feedback from my followers -- those who experience how I use Twitter -- both directly, and indirectly, through their behavior:
- when you go back and forth too much with any one person, your stream has too much activity I may not be interested in, take the conversation on DM or offline
- I follow you because you share much valuable content in a structured approach -- early in the morning when I have thinking time, I like that
- the weekly chat #kaizenblog is the most of your activity and I appreciate that you highlight the content of others, engage in conversation without overwhelming the chat with your tweets, demonstrating you listen and take in the ideas
- love the times when you ask questions as background for a post. I know that you will quote people, and that you will give us an analysis of what you learned, which helps everyone
Twitter is an interesting example, because it's not really ideal for conversation. In fact, the way content flows out of sight and disappears, it's very much a tool that encourages quick bursts, rapid fire messages, and link-sharing that takes you elsewhere for actual conversation.
However the tool works, I can tell you one thing -- don't ever rely on data without understanding the context in which it lives. As you can see in this case, there are many more variables at work than meet the eye. Some data is qualitative and will not be captured in an easy to boil down metric, it has to be experienced to be real.
You can automate everything except your attitude, approach, and people. We're talking big variables here. So beware of falling in love with algorithms.
What's your take? What's been your experience with sentiment analysis tools? Do you cross reference that data with qualitative information?
Weigh in!
© 2010 Valeria Maltoni. All rights reserved.















Hello Valeria,
Great post ! We agree with you: whether your a brand or a twitter
power user like yourself, one should never start with a tool when you develop your conversation/marketing strategy.
The response rate is just one metric in this report and is not an
indication of whether you’re good in conversations or not: as you
point out, dynamics of conversations do not systematically require a reply from you, especially when you’re hosting a chat.
We tried to capture this qualitative aspect in another metric titled:
"You missed". Our 2-day analysis has you at 18% of tweets received which could have potentially required an answer VS the response rate of 45%. Conversation dynamics are complex to capture and we are constantly improving those measure with our beta users. (We would be very happy to improve it with you!)
2 days is indeed a very small dataset and is rarely representative of an overall behaviour (also as indicated in this report, it does not includes DMs).
This report is aimed at, indeed, triggering a conversation and not making a statement.
The overall picture derived from these metrics such as (1) the
proportion of conversation triggers, (2) proportion of personal
tweets, (3) intimacy rate (closeness with your audience) are a strong indication (to us) of someone strongly embracing conversations. Compared to other twitter power users we analysed, you appear to be the conversation master.
We're a conversation management system focused on automatically grouping and segmenting conversations via language analysis. We provide “why people are talking” (25 segments from: agreeing, questioning, suggesting, liking, wanting,....) to enable replies and re-engagement.
What we care is making those conversations possible when volume is high and to provide actionable insights.
Finally, keenkong is a strange name, but it’s a definite conversation trigger and that's what we love about it :)
Olivier
co-founder, ceo, keenkong
@mykeenkong
@olivierberger
Posted by: Olivier Berger | April 12, 2010 at 10:05 AM
Another way to look at this is just to not worry about it. My strategy is to simply use Twitter as an extension of my voice and online persona. I respond when it makes sense to respond. I retweet when I read something valuable that I think my followers will enjoy. I don't really care about my metrics. If I keep up my strategy and provide consistent value to my followers while learning from others, how can I not benefit from Twitter and other SM channels?
Posted by: John McTigue | April 12, 2010 at 10:33 AM
I'd be more interested in seeing a 3 month spread or report like this - I think the nature of Twitter is too changeable and to tell much, normally, over one or two days.
I'm more interested in metrics like this for corporations (based on their business metrics - that's where KK can bring the most value. After all, Eric T. Peterson and bunch of others (ie: Klout.net) have tons of metrics - they're all interesting - though I don't change my behavior after looking at them most of the time.
Anyway, I think the Net Promoter score could be really interesting - though that is one score I didn't talk about since it's still under development.
Posted by: Marshall Sponder | April 12, 2010 at 10:37 AM
A couple of thoughts to add to your excellent post and insightful comments.
1. Quality always matters most in my view. How can any human be intimate and have meaningful conversations with a sphere of thousands (big companies, maybe?... a must if they have a dedicated, talented staff)? Choose wisely for your scale and time, and let go of conversations that are not adding to your life. Kudos for the integrity and candid honesty you consistently inspire through all your media channels. You are a model for how I am trying to approach social media.
2. Blogs for deeper discussions, Twitter for connecting and stimulating dialog. Your #kaizen blog is fabulous. I get a lot out of it each time I'm able to participate and make new discoveries that add to my creative work. More important to me, is what happens here, on your blog. It's a place where your ideas are explored and discussed in more depth. Connecting the conversations between Twitter, blogs, websites, YouTube et al, is key. This is where strategy comes in big time. Being alert to what tool appropriate for what you're trying to accomplish, and how the other tools can work together to help support the idea and discussion, is much more important than metrics alone.
3. Balance. Things are moving so fast, it's a challenge for anyone to keep up with the flow vital information and new ideas. Actively working on engaging in the right mix of media, adapting to change, and adopting new tools in ways that make sense for you and/or your brand — before you spend valuable time — helps keep things real.
Posted by: Peter Rodgers | April 12, 2010 at 11:39 AM
I'd guess the main aim of this was to get you and Marshall blogging about it, so from that perspective it worked well! :)
Posted by: Adam Sherk | April 12, 2010 at 11:53 AM
Valeria, congratulations on your Top 42 Content Marketing Blog award. I always love your posts and learn a lot from you. Blog on!
Posted by: Patsi Krakoff aka The Blog Squad | April 12, 2010 at 06:06 PM
Sorry, I can't but quote Adam's comment: they made you and Marshall talk about the service. Anyway: how do they calculate that a Tweet is worth an answer or not? This I'm curious about the most.
Posted by: gianandrea facchini | April 13, 2010 at 09:41 AM
@Olivier - you did get Marshall and myself to write about it. I'm also fascinated by the fact that we continue to look at tools to tell us about nuances in a conversation we would pick up only if we participated...
@John - indeed! The issue with organizations and metrics is that they can lead to a false sense of security. Can one really participate in social media like they do the rest of marketing, by checking the box?
@Marshall - agree that analysis needs to be based on business KPIs. Know what you measure and why. We have way too much data that is gathered just for the sake of gathering it without a specific goal behind it, nor a commitment to a change in behavior because of the insights you gather from it.
@Peter - thank you for the kind words. I have no choice, I think connectively and do believe it is the responsibility of the person or company initiating the conversation to facilitate it and curate the content wherever it's easiest for people to join. Having said that, I agree on providing a central resource people can refer to. And yes, things are moving fast.
@Patsi - thank you for your kind words. Passion for writing.
@Gianandrea - great question "how do they figure out which question warrants a response"?
Posted by: Valeria Maltoni | April 13, 2010 at 12:51 PM
I've never personally used sentiment analysis tools. I've read and talked to people many times about their inaccuracy. What do you think about them? Have you used them?
Posted by: Cassie Rice | April 26, 2010 at 04:37 PM