6 MINUTE READ | February 24, 2016
MWC 2016 Highlights and Takeaways - Part 1
We were lucky enough to go to Mobile World Congress 2016, hosted in Barcelona, Spain. After a long flight and a few trips the wrong way on the tram, we finally arrived. Here are a few Highlights and Takeaways from our favorite sessions in the first few days:
This keynote was broken up into two sessions and covered mobile content, and ways to use the content to engage users. On average, adults spend an estimated 11 hours a day on their mobile phone, often multi-tasking while also using other devices.
Using content as a way to communicate with friends by sharing stories, pictures, and emotions
If you make something that doesn’t look good on mobile, it most likely won’t be successful or go viral
If users see themselves in it they’re more likely to engage. For example, in a Buzzfeed relationship video, while the content is not the perfect example of a relationship, the user could see elements of their relationship in it and would tag their significant other or friends as a way to communicate
60% of millennials say they’re more likely to engage with a story, something they can participate in, something that can be shared
1/3 of shopping done on black Friday was on mobile devices
Around 41% of millennials use some form of ad blocking already on their mobile device
Using opt-in or push notifications can be very powerful if they’re used the right way and for good, and both the publisher and user accept the use from the start
One question asked by a speaker was ‘could this be the Mobile Century?’. We have all heard the ‘this is the year of mobile’ phrase thrown around every year for the last 10 years. While this is definitely the Mobile Decade, with the amount and type of content currently being created, and the stuff we could only imagine in the future, we’re well on the way to a Mobile Century.
With speakers from Yahoo, Google, AOL, Nestle, and more there was a great in-depth discussion about the right way to approach advertising to users, and ad-blocking. While a large percentage of users already have ad-blocking on their desktop, mobile device ad-blocking is projected to be 0.3% by the end of 2016. We’re in a phase were we’re identifying that there is a problem and can start a discussion on ways to solve users feeling violated by advertisements to the point where they block all adverts.
Through ads, 39% of users find new brands which is huge
It has been found in the last 12 months that 82% of desktop internet users in the UK have some type of ad-blocker enabled on their web browser – incredible!
While larger ads work on desktops, mobile devices don’t have the same network speeds. Slow load times, along with the amount of bandwidth used is a large reason why devices are blocking ads
Due to the amount of ad-blocking, there is now a tradeoff where companies, once providing free content, paid for through ads, are now asking users to pay for their content
With that, a UK study shows 62% of users would still prefer free internet funded by ads vs. paid content
As an ad producer, think about changing the content, placement, and strategy of the ad so that ad-blocking isn’t needed
Respect users by being thoughtful with your ads
Create an experience with more in-depth content, rather than aggressive ads that clutter the screen and slow down the device
Don’t “Spray and Pray”
Instead of “tricks and magic”, use fewer ads that are better quality and in trusted environments
Ad-blocking acts as a blanket, hiding the real problem of bad quality ads
Follow the iab lean initiative guidelines
This was by far one of our favorite sessions! The panel featured Werner Vogels – CTO at Amazon, Dr. Michael Karasick of the IMB Watson Group and Jeff Gehlhaar – VP of R&D at Qualcomm in a very aggressive talk about the new ways machine learning is being used in cars, phones, etc. today. These discussions included different approaches to machine learning and how each panelist differed from the other. For example, Jeff was very proactive of moving machine learning algorithms to the hardware level (SoC) which Qualcomm has chosen, although Werner and Michael aimed at keeping machine learning in the cloud where large amounts of processing power can churn out results quickly. In the end, all the panels decided to “agree to disagree” and admit that different machine learning applications will always differ based on the use-case.
Data from the recent-past can be used to predict the future
Automatically classify categories or prices for new products
Requires good data to be pulled in, garbage data in = garbage data out
Your output data will only be as good as the data you give it. (If you do not trust your input data, you should not trust the output)
The best way to validate the output of your algorithms is to use the 75/25 rule. This means that you should use 75% of your historic data and compare it to 25% of the output data to see if your model is correct
Be smarter about what works on each device. Mobile should act as a window into data, with the core functionality running in the back end/cloud. Mobile is limited by the size of the dataset. A mobile device will not be able to process large amounts of data efficiently.
Machine learning is simple. Algorithms designed to compute data. The difficult part is analyzing and knowing what to do WITH the data.
It’s always about the use case. Machine learning isn’t for everything. (Ex: Health related opinions, etc.)
Allowing computers to make decisions for us will not allow them to take over the world. They are simply computing answers for the tasks that we instruct them to do. (#skynet)
At PMG, we have been experimenting with machine learning and how it can benefit the ad industry and found the advice given by the panel to be extremely helpful and eye opening.
Stay in touch
Subscribe to our newsletter