7 Mistakes To Avoid when Selecting a Measurement Partner for Social Media
New York, NY, September 25, 2017
Written: Tania Yuki, Founder & CEO, Shareablee
Source: Linkedin Article
1) Looking for a Platform that can ‘do it all’
Would you willingly choose a restaurant that claimed to have the best bolognese, burritos AND sushi in town? Probably not – you know all too well that one restaurant can’t be great at all those things at the same time. Social media works the same way – if the platform can schedule posts, optimize your ads, plan and buy your media, listen effectively in real time to customer sentiment AND benchmark your performance against every relevant brand in your market, it’s probably mediocre at all but one to two of these functions… and maybe all, as it’s distracting to pursue loads of highly different disciplines in the one technology and product team.
While it may seem less convenient, you’ll get much more for your investment if you take the time to pick a partner committed to being the best in the world at one core skill. But make sure that partner has great APIs, as it will be highly valuable down the road to integrate into centralized dashboards with all your other key marketing metrics.
2) The “Good Enough” Mistake – Picking a Tool that Meets just your Maslow’s Basement-of-Survival Needs
Our industry and the world at large is experiencing the most fast and frequent change in history across media, retail, science, technology and many more areas. As Thomas Friedman puts it ‘the world is not just radically changing, it is being dramatically reshaped to operate differently in many realms all at once. And this reshaping is happening faster than we have yet been able to reshape ourselves, our leadership, our institutions, our society…’
If the partner you’re working with cannot provide a compelling roadmap into the various futures out there, then they’re already obsolete – and taking you with them. As painful and expensive as it is, each partner you work with has to be actively anticipating the future, and it’s well worth understanding if this is the case prior to making a year long, or multi year commitment.
3) Choosing a Partner who Doesn’t Measure the Whole Terrain
When I founded Shareablee in 2013 I swore blue that we would not build a dictionary of all brands in the world from scratch – the thought sounded simply too painful. Instead, we rolled out measuring You, and whoever it was you nominated as your ‘competitive set’. But very quickly, it became clear to us that this would not work for the future – the future, in which who you believe you compete with may miss who’s really about to disrupt you, and where competition is not only local but global and not only from within your own existing industry, but beyond it. Today, we track almost all global brand pages and this number grows every day as the influencer and digital-first ecosystem continues to evolve.
Picking a partner who is not proactively monitoring everyone out there who may deliberately or accidentally be trying to kill your business, simply doesn’t cut it anymore. The world around each brand is changing too fast, and it’s too dangerous to not be able to know what’s coming.
4) Not Knowing what you don’t know when it Comes to Methodology
Knowing which data points ARE NOT available is just as critical as knowing what you must have, because there is a lot of poorly created information that will gravely mislead your business (eg Video views estimated based off of your engagement – for no defensible reason).
Here is a cheat sheet for which metrics are not available on social, which you must be cautious about if promised- each provider who claims to have these are going about creating them in different ways, which may or may not be remotely accurate:
(how are these calculated? Good answers will include taking into account both pass-through insights data as well as machine learning models/other estimation processes)
– Uniques/Unique Engaged Audiences
(Good answers will include taking in census level activity, and de-duplicating people deterministically on that basis)
(No real good answers for this, other than models built off of the above metric that seem plausible. If the provider does not collect census level audience data, there is absolutely no way to get at this metric other than passing through your own insights data)
– Video Views
(Are these modeled or scraped? There is currently no video data directly from Facebook, Twitter and Instagram other than the video view data available for publishers from Facebook themselves, via CrowdTangle)
– Paid Media Identification
(How is this identified? Good answers will take into account both actual as well as modeled data, and will do more than merely flag high performing posts as ‘promoted’)
(How is this identified? This one is really tricky – plausible answers will have a very well defined methodology, so be very skeptical of any probabilistic modeling here)
– Industry Averages on any metric
(How are you defining ‘industry’? How do you know you measure everyone and how are your industry categories updated? From experience, it takes both technology and an excellent taxonomy team to maintain this as industries change in an ongoing way).
5) Getting stuck in the details of micro, highly customized metrics (“aka if you can tell me my hidden unliked-likes from just re-shared, dark social on Thursdays in Cincinnati, I will be happy”)
This is my personal favorite – and the analyst in me loves each and every person guilty of this one. Shareablee has over 800 metrics available now in our dashboard (of which probably about 50 of these metrics are heavily used across the industry) – which means we cater to a lot of nuances and edge cases. We were after all built by researchers, and therefore are prone to think through and create metrics for a lot of randomness.
That notwithstanding, still we will occasionally work with a partner who is completely committed to metrics that don’t exist, that they have absolute certainty that they requirefor their business or else they can’t even get started. By the way – the metric might be wonderful and ideal, or exactly something that they’ve used measuring other formats (eg GRPs in television). I’m not discounting the thinking. My observation here is: don’t let perfect get in the way of possible. Just get started. The future of research and of analytics is not to have perfectly calibrated, small sample data that isn’t messy but rather, to meld larger more disparate datasets together to make meaning. Think through with your partner what you’re trying to understand with this metric, and the right partner will either be able to honestly tell you it’s not possible right now, or point you towards metrics that can help you get the insights you’re looking for, albeit gleaned via different metrics or perspectives.
6) Choosing Form Over Function
Everyone loves beautiful charts, and ideally your provider will have both beautiful charts, and best-in-class data. When dazzled with gorgeous charts and dashboards, here are a few practical questions to ask:
– Can I easily export the data that’s relevant to my business, in a format that I, or my analytics team can use?
– Are any downloaded charts editable (that is, loaded with real data) so I can customize my view
– Does this provider have proprietary key metrics, or are they merely visualizing data that I already have (albeit in a more attractive way)?
It is easy to be momentarily duped into thinking you’re looking at data that is special and actionable, when it is visualized beautifully and clearly. But the insights won’t be there.
7) Blindly trusting in ‘Predictive Analytics’ to deliver ‘Actionable Insights’
I know – there’s a crap ton of data out there and it can be exhausting to even think about thinking about how to ask the right questions of it. But that’s unfortunately the rub these days. Solutions that claim to automate the thinking right out of the process of analytics are dangerous at best, and telling outright fibs, at worst. Just like buying a treadmill won’t ensure that you get fit, merely owning a great solution has no bearing on whether you’ve set up your team and culture to actually use it effectively. Analytics needs your best thinking now, more than ever.
Please don’t misunderstand, I’m a huge proponent of predictive analytics and their use case when it comes to deriving insights from big data. ‘Big data’, or whatever you want to call it, is getting bigger by the second, just as what is measurable is only getting bigger and it’s not reasonable for humans to process it all with our limited CPUs. But when a partner claims to be able to ‘predict’ what to post, or to ‘automate’ your analytics, make sure you dive into why they believe they can do that for you. Most ‘predictive’ metrics are really descriptive metrics, and as noted earlier one partners’ solution to ‘predict’ paid media is to mark it as ‘paid’ whenever it’s over that page’s average (which by the way, is grossly inaccurate).
Wish to know more?
Join our State of Social Media in the Beauty Industry Webinar 2017, where Tania Yuki, Shareablee Founder & CEO will be unpacking how emerging and established brands build desire, trust and purchase for the fully connected consumer.
Click: Reserve your spot today