We live in a world driven by data and although we probably don’t even understand half of it or don’t even bother to look into it as we should, decisions aren’t taken unless they can be fully rationalized. As Rishad Tobaccowala mentioned in a recent post:
“In marketing we worship the algorithm and its superiority to human decision making.”
He makes a good point. He continues with:
“In the world of media we are so fixated on the plumbing of finding the right person at the right place at the right time that we forget that the interaction we deliver will have to be absolutely right and brilliant not to piss of this superbly well located person at the exact right time. The better the “targeting”, the more important the tone, content and quality of the interaction. Lets think about the poetry versus just the plumbing.”
This reminded me of the conversation between Bill and Melinda Gates during this year’s TED event in Vancouver. Somewhere in this conversation – hosted by Chris Anderson – it’s clear that part of the magic between these 2 people in spending billions of dollars to charity is the mathematical approach of Bill Gates combined with the more tangible, human experience of Melinda with the people involved in the decision. Something they obviously recognize as a necessity in their decision making.
There’s a lot of talk about shareable content and what it means to go viral. This post is in no means a guide to make sure content does go viral, with this post I hope to help people understand why content spreads and if that is really what has happened in the first place. The reason I think this matters is because marketers are judging other brand’s content to understand what worked and what didn’t hoping to replicate, but they aren’t necessarily always looking at the right data. And since I get into conversations related to this topic constantly, I figured it would be well worth sharing this idea of share rate here as well.
Enough for the introduction, I will use online video (ads) as an example, but the theory works for all kinds of online content, although the data sources will be different of course. There are two key numbers that people tend to look at when judging the succes of online video: views/impressions and shares, although I presume that’s already a distant second measure.
Youtube for instance has its own way to rate popularity of video ads, that is explained by themselves rated based on “an algorythm that factors in paid views, organic views and audience retention” – the Youtube Leaderboard. It’s unclear what the weight of each element is, but it’s clear that it’s mostly based around views. It is also the only number mentioned in the rating. As an outsider it’s difficult to judge only on that number what made each video succesful, was it the idea or was it the mediaspent? Knowing what Youtube’s business is about, there’s no need to explain why this rating makes sense for them.
So it’s important to look further. If you as a marketer (or agency creative) want to figure out why something worked views aren’t the best number to look at on its own. Unruly Media created another way to measure video by looking at the amount of time something was shared on social media. The ads chart is sponsored by Mashable, but you can also look at popularity of other video content. If you want to understand why content was spread amongst people it’s probably a good idea to check if it actually spread in the first place.
Here’s where it becomes interesting. The ranking based on shares looks pretty different than the one based on views, if you look at Youtube’s n2 for August for instance, you will see it’s almost impossible to find in the Unruly ranking. So it’s clear, you wonder how content spread? Look at the shares. But that’s not all.
Let’s look at a classic video we all know for instance: Evian Babies. With over 3 million shares that puts it at number 7 in the Unruly Viral Ads Chart of all time. Very succesful, but does that proof it was spread across social and hence a big succes? Not quite. Paid views will also generate shares – paid and organic. So we need to look beyond that. Here’s where the share rate comes in. The best way to judge whether a video was viewed because people shared it across the web is to look at the ration between views and shares. There’s a few ways to look at it, we use [shares/views] as the ratio, some use percentages, the idea remains the same.
Truly viral content such as ‘Dumb Ways to Die‘ or our own ‘Push to add drama‘ will have a share rate in between 1/20 and 1/10 or even higher (meaning 1/9 or 1/8 but you won’t see those number appear much). Evian Babies – to come back to that same reference – has a ratio of 1/40 and the other example I mentioned (Foot Locker – Youtube’s August n2) has a ratio of 1/200. I would reckon that everything below 1/15 (probably) or 1/20 (definitely) received some kind of ad push. The lower the number the more views were generated through advertising (versus organic) obviously.
The share rate on its own doesn’t say much either of course, a high rate with little views isn’t much of a succes. But if you really want to know why a video was seen by so many people then this is the measure you need to look at. Does that mean those other videos weren’t successful? Of course not. Is it wrong the push videos online with media? Ofcourse not. But want to understand where success came from with the little data you can access? Find the video on the Unruly chart (they show both views and shares – makes it easy) and calculate the share rate.
For the record, Youtube also has a kind of share ratio they use in their presentations but it’s not meaning the same thing. They will look at the ratio between paid views and organic views. Again, thinking of their business selling video ads, I makes for them to correlate paid versus organic views, rather than views versus shares as I suggested.
As mentioned in the beginning I used video to explain but this idea of share ratio counts for all types of content and thus should help you analyze success (or not) of others in good way.
(Please note that I only use examples to illustrate a point, it’s no judgement at all about the videos themselves).
I found this video from the interactive prototype Room-E on the 72U project blog. It’s a prototype showing what will be possible in the near future when you think about a more responsive environment.
“The future of the computer is to essentially make it disappear—a disconnected interface, so the house or the office or the building or the city is the computer.” —Mark Rolston, Chief Creative Officer, frog
Alright. There’s not much I will say about this, you just have to watch the video. In short, the organisation Terre Des Hommes that fights child exploitation, created a robot that looks like a 10 year old child. This robot, called Sweetie, is operated from Amsterdam and once online engages in chats with pedophiles. Apparently when you go online on popular chat services with the profile of a 10 year old Philippine you attract these sex offenders within seconds so that’s what Sweetie’s for. And since they all ask to put on the webcam, Sweetie activates that webcam without any hesitation… and while the conversation lasts, the specialists in Amsterdam get photo & video evidence of the offenders and they try to find all information that helps identify these men. And it works: 1.000 pedophiles identified in merely 2 months. I don’t say this often but this is just amazing! Watch. And don’t forget to sign the petition.
Big data is the whole grail of marketing. And yet not many is actually making lots of progress. There’s a good good on that captures well what the state of big data is if you ask me:
“Big Data is like teenage sex. Everybody talks about it. Few do it and they do it in the dark”
As always also here there are the exceptions that prove the rule. Companies like Starbucks or Taco Bells have showcased that they are actually using data to the extreme to help improve their business activities and communications. But in general only little data is being used, it’s something we encounter only occasionally during the marketing activities and also as a consumer I don’t see much return on the fact that people seem to know quite a bit about me.
And I wonder if that isn’t going to become a problem sometime soon. As a tech savvy consumer I know that companies have data on my consumption behaviour, I am very aware whenever I need to give someone personal data and I know that a lot of my online behaviour is public for everyone to see. Because I know, I also expect something in return. There’s a very one on one relationship between filling in a form before being able to proceed to a next stage and in that case you can immediately judge whether handing over that bit of data was worth the return. But that counts also on general consumption, on the data that companies can gather by tracking your behaviour. I know they do and also there, even if less one on one, I expect a return.
I’m ok with my data being used, I want my data being used or even more, I demand my data being used. I think consumers will get ever more aware of the fact that their data is being collected and as a result become more demanding on how they are treated. No more useless questions, seemingly random suggestions, repetitive data collection, … And that makes in my opinion the need for companies to start really using their ‘big data’ even more important. Not just because they indeed can improve their business and communications if they use it right, but because they have to as consumers will start demanding that. So big data is great, but it’s also a big promise!
So if you’re a business owner or a marketing specialist I think you really need to start figuring out which data you currently gather and how you can connect that to make consumers lives better. All in all I believe there are 3 kinds of data we have to think of that are key to improving your marketing & sales efforts:
Form fill: data that you’ve asked consumers to fill in, can be at various types of touchpoints, all data people hand over to you to get something in return
Usage: usage/consumption data from both sales as marketing activities, everything you collect through the customer journey
Public domain: everything people share in public online & offline that relates direct or indirect to your business
So think about it. The companies that are using it are outperforming you and taking a lead and your customers will move to those companies because they are living up to the data promise making the gap even better.
Hyperlocal is a hot topic these days. Especially for news media it seems to be where the future lies. You can either try to play a global role like The Guardian or Al Jazeera are doing with success, or find a way to be more locally relevant. And I believe the future of news for most publishers will be more local than it will be global. Because let’s face it, only when you publish in one of the world’s most spoken languages you can make a difference on a global scale.
The problem with hyperlocal is that most people interpret it wrong. Currently the most common practice is to use geolocation, you either enter your postal code or the application/website defines your location based on GPS or IP settings. So you get news, offers, information related to where you are. Very efficient but when thinking about hyperlocal the opportunities are beyond that.
The reason why is the following:
I got many localities that are relevant for me. I think about the place where I live, where I work, where I lived when I was young, … I have a bond with all these locations in a certain way so I’m interested in news from all these places and even other types of information. Today no-one is offering me a way to organize news gathering with this in mind. At best, like I mentioned before, you get that output for one of the chose locations, never for a mix of. Prove me otherwise.
The range of interest centered around me can differ in size depending on the topic and the moment. What I mean with that is that when I’m looking to buy new furniture I’m probably okay with driving a bit further than when I want to get a pack of cigarettes, just to name something. Or when I want to buy a car versus when I’m looking for an accountant. This means for me the range you take into account when talking about hyperlocal needs to be quite a bit more subtle than it is today.
Last but not least, hyperlocal doesn’t mean faits divers by default. The local version of my online newspaper tells me about a selection of the tiniest events I don’t care about in my city, but fails to mention when the Tour de France is passing by. Zooming in on hyperlocal seems to mean the same as zooming in on stuff that nobody cares about happening 2 blocks away. Why is that?
With more and more information, media being geotagged and with technologies that allow geofencing I hope that we’ll see an evolution in the way we deal with hyperlocal that takes the reasoning I wrote down into account.
Fascinating. And visually attractive. The people of LinkedIn Labs just recently created this InMaps application, a kind of analytics tool to “visualize your professional network, clustered in realtime based on their inter-relationships”. A pretty cool tool actually, and I’m a sucker for these kind of applications.
Log in with LinkedIn and the tool will analyze your network and visualize it in a graph like the one below, which is actually the output of my LinkedIn network.
What’s extra interesting about it is that the output is dynamic (unlike this image) and that you can hover over each contact to see their specific connections within your network. That way you also get a view of how the clusters are made and InMaps allows you to put a label on each colored cluster to make it easier to see who’s who. Just give it a try, you’ll see for yourself.
Interesting results for myself is to see for instance that I have 2 Microsoft clusters (I’m ex-Microsoft remember), one for MSN/Windows Live related contacts and one for more general Microsoft contacts. Interesting to see that this split is made, although it’s actually pretty logical when you look at it. Also interesting is to see between which groups exist more links, not always what you would expect. I’m definitely not done analyzing this, but curious what your graph/learnings look like so please do share ;)
Last but not least, it’s also pretty interesting proof that people are organized in groups, clusters and that if you want to influence people it’s important you understand these clusters – or ‘spheres of influence’ like we used to call them, dixit David Armano.
In February of this year I wrote a post about Kevin Slavin’s talk on Augmented Reality at PICNIC NY Salon. In that video he talked about something that made total sense to me… which to be honest is true for most of what Kevin says anyway :)
“His thoughts around augmented cities and why maybe ‘augmented’ should be about taking things away instead of just adding them to the world as we are already drowning in data as it is.”
So when I got this video today from a colleague about a research project on ‘Dimished Reality’ by Jan Herling and Wolfgang Broll of the Ilmenau University of Technology, it was like a proof of the concept Kevin talked about a year ago now. I don’t like the name ‘Dimished Reality’ because it still is doing more on top of what is really there. But in this case less really is more, check it out:
The YTTM offers an interesting way to watch videos from a specific year in between 1860 and 2010. Pick a year and choose one or more categories (video games, television, commercials, …) and you get a video that fits the selection.
Something bugs me. Not a day goes by or new usage data (preferably in the form of an infographic) gets shared online about one of the favorite social media initiatives such as Facebook, Twitter, … you know the lot. Big data, big numbers most of the time. What I don’t get though is why we all seem to copy/paste most of that information on our own blogs without really trying to understand what the numbers tell us (and what they don’t tell us). Everybody who once worked in a PR related job knows that companies publish numbers in a way so they look good. They use absolute numbers when they are worth it, percentages when they don’t look good and so on and so forth. When I say visitors to this website using Android have doubled over the last week (+100%) that is sounds much better than if I were to say there are now 2 people using Android to visit this blog instead of one. You catch my drift, I would really like to see some more analysis on those numbers before publishing if that’s not too much too ask.
Something else bugs me even more. When making these ‘analysis’, infographics and what not, people are not comparing apples with apples. Nobody seems to find it a problem that we’re always comparing 500M Facebook users versus 145M Twitter users (and some even against the 300M Windows Live users). For Facebook that are registered users, and as such most likely unique users. For Twitter that are registered users, and most likely that means registered accounts – and not unique users. I’ve got one Facebook profile just like most people but do use 3 Twitter accounts (@crossthebreeze, @iblogmustang and @krishoet). For Windows Live however the 300M users mentioned are active users, active meaning that they’ve logged on to the service at least once during the last 30 days. You can discuss about whether that is a good measure for being active or not, the point I want to make is that although they’re all big numbers they all don’t really mean the same thing. And that makes it unfair to just compare them like they are in my point of view.
Especially the registered versus active users is something really important to think about. When promoting webservices such as the ones we’re talking about you can imagine that generating awareness is the first big task on the agenda just like any other company. But because they are webservices I presume once you get the attention needed, driving registrations is not the toughest part. Registering to an online service is easy, I’ve registered to hundreds of services by now but use only a percentage of those on a regular basis. Activating users/consumers is the toughest part. People show interest when the buzz is up, but what is it that you do to keep them interested? That’s a tough challenge, a challenge to which many services fail if you ask me.
And it’s not just webservices of course, same counts for apps etc. There’s a boatload of apps available for my phone apparently and still I find it hard to find a dozen decent ones to download on the device. So don’t just report on the big numers PR people give you, those don’t always mean much (at least not to me). And please compare numbers worth comparing, otherwise that makes no sense either.