4. Data merchants =
data overload
Consequences:
− Confuse correlation and
causation
− Dump data without
context
− Lazy filtering
5.
6. Data merchants =
data overload
Consequences:
− Confuse correlation and
causation
− Dump data without
context
− Lazy filtering
7. OUTPUTS OUTCOMES
NUMBER OF
ARTICLES/
REACH
VIEWS
ATTITUDES BEHAVIOUR
BLOGS FORWARDS BELIEFS PURCHASE
PERCEPTION
TWEETS RETWEETS
TONE
SOV CONTENT
COMMENTS
CONSIDERATION
PREFERENCE
RELATIONSHIPS
(NOT A ONE
NIGHT STAND)
LIKES
8.
9. In 140 characters or less
Data merchants collect more than
standard identifiers of a personal profile.
This is an opportunity and a challenge
#internetshowmel
About 2 minutes ago via The Internet Show | Social Media World 2012
11. a∙dap∙tive mar·ket·ing [uh-dap-tiv mahr-ki-ting]
An approach that allows marketers to tailor their
activities in rapid or unparalleled ways to meet their
customers’ interests and needs based on data.
13. Measuring the
What to measure:
the basics sales funnel
Research:
Share of brand
New key search phrases
Digital reach
Depth of interactions Opt in audience
Share of voice Planning:
Depth of interactions
Share of voice
Key category phrases (buckets)
Traffic by media mix
New key search phrases
Testimonials:
Traffic by media mix Depth of interactions
Share of voice
Digital reach
Digital reach
Lead capture/sales?
Opt in audience
14. Best practice
What to measure:
the gold standard reporting
Set up keyword queries with your audience in
mind – search insights can help you with
common spelling errors or relevant terms
Volume Never trust automatic sentiment – the
Australian lexicon alone is enough to give
false results (terms like “not bad”, sarcasm,
Events sick/mad)
Triggers
Conversation clouds
Sentiment (manually filtered)
Top authors
Top domains
17. In 140 characters or less
It may be relevant to compare Apples and
Red Bulls. Set benchmarks against social
objectives AND product categories
#internetshowmel
About 2 minutes ago via The Internet Show | Social Media World 2012
21. Writing your report
− Clarify your objectives and key performance indicators
− Present the numbers
− Write an analysis and add context to the numbers
− Highlight the changes from previous reports
− Include recommendations
− Include anecdotal feedback
− Ask for internal feedback
24. Free Tools
Find some help
Website Audit
http://delicious.com/mab397/websiteaudit
Social Audit
http://delicious.com/mab397/socialaudit
Platform Search
http://delicious.com/mab397/platformsearch
Search Audit
http://delicious.com/mab397/searchaudit
Alerts
http://delicious.com/mab397/alerts
25. In 140 characters or less
Great reads: Occam's Razor by @avinash
http://bit.ly/Ijvya0 & @thebrandbuilder's
Social Media ROI http://slidesha.re/IjvJCa
#internetshowmel
About 2 minutes ago via The Internet Show | Social Media World 2012
26. GET IN TOUCH
CONTACT US
Need to discuss this in more detail?
Whether you want to clarify our approach to
strategy or just have a burning question about
digital in general, please contact me at any time.
And feel free to stalk me online as well – here’s a
few places you can start.
twitter.com/digitalmands
Address:
au.linkedin.com/in/mandibateson Level 15, 65 Berry Street
North Sydney NSW 2060
delicious.com/mab397
Phone:
02 9287 8505
mab397.wordpress.com
Email:
twitter.com/mab397 mandi.bateson@mindshareworld.com
26
27. Credits
Thank you to these photographers for their images
Slide
3 JD Hancock | Flickr
10 Fetish Art Identification | Flickr
12 Heavy Weight Geek | Flickr
15 King_Of_Games | Flickr
22 Pedrosimoes7 | Flickr
Editor's Notes
Applications are collecting data from us beyond the standard identifiers of a personal profile. Data collection goes beyond name, DOB and email address.If you’re not paying for it, you’re not the customer. You’re the product being sold.Geo social games and recommendation enginesKnow not only if we’ve delivered an impression but if they like it, if they saw it because a friend recommended it, if they shared it with their friends, how they identify themselves (tags or categories – food, fashion, melbourne, F1)This puts us in data overload and the consequences include:The confusion of correlation and causation. If you’ve grown an opt in audience it doesn’t necessarily mean this is a true representation of your audience. Pinterest followers now would be early adopters, up until a fortnight ago Instagram would have been only iPhone users. It makes it easy to dump data without context – Facebook has external measurement tools that release details of the most engaging pages – they haven’t taken into account whether a page has a high number of comments or wall posts because it runs it’s page like a customer service channel. It doesn’t even look at the sentiment of the comments. It’s too easy to see the percentages churned out by tools and taking them at face value. Social monitoring tools can scrape so much content across the web that even if you use some very specific and detail keyword queries you still need to sort through all responses to sort the data to suit your reporting. What mentions do you want to track? Insults vs product mentions (health management client)What makes a positive or neutral mention?Filtering brings further insights – an airline like Qantas could tag mentions by relating to service, the frequent flyer program, the corporate performance and today industrial action and once all mentions are categorised you can improve the way you view your dataUnderstand that this is both an opportunity and a challenge.
Context – here we see a range of comments on Twitter regarding the NSW election – people are telling us when they’re voting, they’re questioning politicians directly, they’re making glib comments, they’re sharing moments from the election trail. Beware of influencer ranking tools – many qualified those who had been retweeted regularly in elections as influential. There’s an obvious difference between Wil Anderson cracking a joke that you are likely to RT versus his influence on your voting preferences and the automatic tools don’t take the time to qualify this.
Applications are collecting data from us beyond the standard identifiers of a personal profile. Data collection goes beyond name, DOB and email address.If you’re not paying for it, you’re not the customer. You’re the product being sold.Geo social games and recommendation enginesKnow not only if we’ve delivered an impression but if they like it, if they saw it because a friend recommended it, if they shared it with their friends, how they identify themselves (tags or categories – food, fashion, melbourne, F1)This puts us in data overload and the consequences include:The confusion of correlation and causation. If you’ve grown an opt in audience it doesn’t necessarily mean this is a true representation of your audience. Pinterest followers now would be early adopters, up until a fortnight ago Instagram would have been only iPhone users. It makes it easy to dump data without context – Facebook has external measurement tools that release details of the most engaging pages – they haven’t taken into account whether a page has a high number of comments or wall posts because it runs it’s page like a customer service channel. It doesn’t even look at the sentiment of the comments. It’s too easy to see the percentages churned out by tools and taking them at face value. Social monitoring tools can scrape so much content across the web that even if you use some very specific and detail keyword queries you still need to sort through all responses to sort the data to suit your reporting. What mentions do you want to track? Insults vs product mentions (health management client)What makes a positive or neutral mention?Filtering brings further insights – an airline like Qantas could tag mentions by relating to service, the frequent flyer program, the corporate performance and today industrial action and once all mentions are categorised you can improve the way you view your dataUnderstand that this is both an opportunity and a challenge.
We tend to focus on the OUTPUTS. What actually happened. The number of articles, blogs, tweets. The total reach of those, the number of forwards and retweets.What our SOV was and what the content and tone of all that content was.This makes us feel good. It is easy to understand. And relatively easy to collect. And here size does matter.But does it? Are numbers the most important thing?Would you rather reach 1000 of your customers with a fairly run of the mill communication, or 100 from a really influential source that convinced them all to go buy your product? Obviously the later.The harder part is linking these to OUTCOMESWhat was the impact on consumers attitudes & beliefs – and ultimately their behaviour. If you’re willing to invest the time and the budget, you can use social media monitoring to determine these outcomes, whether a campaign was online or offline. You can see buckets of phrases when people mention your brand or product, like volcano – spectacular or dangerous. You can see whether people talk about your product on a financial level (I bought this, it was on sale, it was cheap, it was expensive) or idenitfy other “emotional” triggers (I drank this, I’m about to eat this, I just did this and I feel great, here’s a photo of my new purchase do you like it?). We use this to help inform future creative – how people see themselves and how we can relate to them with advertising.
An overload of data is a great problem to have. And most of the issues can be solved if you do one thing – keep going back to your metrics and ask yourself why?
Mindshare’s global digital lead Norm Johnston’s definition of adaptive marketing. Whether it’s optimising Facebook advertising performance or using engagement rates to determine how to write the best call to action, the important point here is not to run campaign reports at the 3 or 6 month mark and feel like you’ve measured your success. Jesse from Tourism Australia showed us yesterday that he can tell what colours will provoke more of a response on Instagram when he posts images. RT insights are being rolled out in Facebook and even with 48-72 hour delays you can still learn so much by continually evaluating your campaign.
Look good amongst your competitors and stand out from the crowd
You can track your performance against competitors but make sure you’re playing on the same field. Each company is using their social networks to achieve their own strategy which may mean it’s not right to compare one Twitter account with another or one Facebook page with another. Be critical of your benchmarks – if you are using social to make a connection with your audience and strengthen your branding, look at comparing your performance outside of your product category
Some people just want to win awards or get 50,000 new Facebook likes. If your stakeholders have these kinds of goals that can have an impact on your
Social media gives you the opportunity to take both a micro and macro view of the overall business objective. Overall business objective is to be profitable and successful. Secondary objectives are to hire and retain great staff. A company goal may be to become an employee of choice. Companies may investigate the success of unified communications, green buildings, hot desking, the effects of commuting on a happy employee – all these other areas of consideration that may have an impact over the physical location of their business.