Thu, 26 Jul 2018 23:12:19 GMT
In the latter part of 1971 Ray Tomlinson sent the first email. Since then telecommunications and computer networks have become increasingly complex and powerful. Today every aspect of our lives relies on fibre optics and data centers. IBM estimates that every day the world generates 2.5 Exabytes of data, for context that equivalent to streaming 4K video for over 40,000 years.
This hyperbolic trajectory will likely continue, but today progress is less defined by how much data we generate, how fast our phones run or our connection speed and more about how well you can filter the noise. Google, Facebook and Amazon are all proof of this; they have all built their empire on taking the firehose of the internet and making it usable.
I’m declaring the information age as over; we’ve been in the filtering era for some time. In recent years agencies have been focused on moving their operations piece-by-piece into the ‘cloud’, replicating their existing structures on remote servers. I’d argue given the sheer volume of information being generated, upgrading systems, while necessary, shouldn’t be considered as progress. Progress is being able filter and leverage data.
So with this in mind if you were building an agency to make the most of the last 50 years of progress from scratch and didn’t have the baggage of an existing organisation how would you do it?
The answer to that question would need a book (or two) to cover, there are so many considerations, but here are four near the top of my list:
1. The 5 Vs of Data
If you’re collecting data to improve your client’s experience and service you’d do well to consider the 5 Vs of data:
How much data should you collect, what amount do you need give you meaningful information?
How many different kinds of data do you actually need / from how many channels (and how much of it is actually useful)?
Is your data accurate? (Are you sure?)
How fast is your customer data travelling? (what rate are your customers expecting the data to be reflected in your product?)
Make sure that you are creating more value for your customers than you are capturing.
2. Security is everyone's responsibility
Macy’s, Adidas, Sears, Delta, Kmart, Under Armour, I could go on…. All have had data breaches in the last year. While you’re working on making your agency future proof don’t forget to consider security. Computer security is hard - really hard - and the bigger you are the bigger the target but you can protect your and your client’s data.
With some basic best practice rules you can operate securely on the cloud, your biggest challenge is not the technology but compliance within your organisation. Placing most of your effort on educating your staff and driving best practice across your business will help protect you. We rely every day on computers, it’s important to make computer security everyone's responsibility. If you’re using solid encryption at rest and in transit, even if data is exfiltrated it won’t be readable.
3. Serverless and Containerisation
Moving to the cloud and managing data and services remotely used to mean renting servers, that era is coming to a close. Nowadays it’s more efficient, less hassle and more reliable to do away with the servers and run every request as its own function. Applications are ‘containerised’ so that a cloud platform like Amazon Web Services can take the container and run it’s contents automatically without the need for you to deploy it to a server. This trend is yielding best practice methods that offer high availability and powerful computing horsepower in a much more affordable way.
4. Understand what machine learning is… and isn’t
The words AI and Machine Learning get bandied about all the time. Many people imagine AI and machine learning as an all powerful super technology that can be thrown into any situation to offload tasks and automate everything. We’re not there yet, and I would suggest using the term AI anywhere is premature, the terms ‘machine learning’ and ‘deep neural networks’ are far less hyperbolic and more accurate.
So what can machine learning offer agencies today? There are many applications of neural networks, I believe the most relevant for agencies and creatives is adaptive content delivery where a algorithm watches users and delivers content and user experiences based on their behaviour. There are some companies already working on this in the context of website performance and of course platforms like YouTube rely heavily on these algorithms to serve content. In a world where people are increasingly immune to traditional advertising and content marketing is proving hugely effective, agencies would do well to consider how they can leverage machine learning to serve more relevant messages to their client’s audience.
What machine learning can’t do (at least yet) is sanitise data, the old principal of garbage in garbage out still applies, enter point 1 and the five Vs of data.
Charlie Cooper is product manager at Blutui, Blutui.com the web platform built for the creative industry.
view more - Trends and InsightBlutui, Thu, 26 Jul 2018 23:12:19 GMT