SMX London 2013 was opened by Jim Sterne who focused on a key role for the search professional as a company insight wizard.
He started by explaining the role of the company insight wizard. The focus needs to be on moving the company forward and not specific tactics. Search professionals have their hands on the numbers and you can use this to be a great analyst.
Understand the raw material
Data is immense and very difficult to visualise. Volume is more than we are accustomed to and it is exceptionally varied. The amount of information is incredible from keywords to third party data which results in data overload. You have loads of touchpoints and each of these tools have different reporting. What we need to do is pull all the apples and oranges out and get it into a single system to analyse.
You are responsible to manage that data in that you need to trust the data and what it tells you. You need to understand the data and question it when it needs to be questioned.
Understand the tools
Yahoo and Google were the first pioneers to manage big data and they did that by de-centralising it from the original mainframe. They did the same with pre-processing and parallel processing. Big data was developed in that time and allowed us to do great data visualisation stuff! The point is though now you don’t need to know the question before you start storing data. Now you simply need to store the data and then analyse it in the future. There are a plethora of tools from Experian to Google Analytics to Crazyegg who all collect the data you may need. You have so much data and so many tools that its almost useless because they don’t provide insight to the problem.
Understand the problem to be solved
Traditional data collection is solving the wrong problem. To identify the problem you need to:
The only three meaningful goals are:
Understand the art of analysis
From here we have a lot of data and a lot of tools and a problem to solve.
The low hanging fruit:
Look at the anomalies – check for errors and omissions, complaints, spikes and troughs and oddities to identify opportunities. The pops and drops make the different.
Segmentation is your friend – it almost doesn’t matter how you segment as long as you do. You need to segment to make sure you don’t believe in averages. Look at the micro conversion paths to identify the major drop offs. It’s really important though to realise that the issues are further upstream.
Correlation is not causation – the towns where there are more churches have more alcoholics but it isn’t causative the population size is. He also showed a chart where murder rate and internet explorer usage is almost exact.
Understand the art of marketing analysis
He then shared an example of trying to improve the lifetime value of a customer and showed visual representations of analysing the key customer. He was left with the money buckets and then we can start segmenting them to understand the marketing techniques that helped him win more of them. We then start testing ideas across those segments and measuring the success. To be a great analyst you need to understand:
How do you get better at analysis?
Intuition is so important as you keep getting knowledge all the time but you get the “aha” moment when you are most relaxed and are not thinking about the problem.
Reporting is what you end up doing but it’s not your best product. Analysis is your strongest skill. Data are grains of sand and fascinating but you need to understand the wind that moves the dune.
At the end of the day it is about asking good questions is the secret to being a great analyst.
The art of communication
Tell stories instead of delivering reports. – he shared a story about an accounting department that stopped sending reports in a company and a single person asked for one out of the vast array of reports they were sending. Figure out how to tell stories.
Don’t tell people how hard your job is. Make recommendations not excuses. It’s not impressive that you had to work the weekend the insights are.
Tie it into individual’s goals. Each person wants something personally and help them achieve those goals.
Have an opinion. Tell people what you think in a way that allows them to action something on it.
Be specific. We can improve this number by this percent in this time frame with this investment and we can prove it with this test. Don’t show people the numbers necessarily tell them the story about your opinion. It’s not bean counting, its percentages and probabilities.