How the codebreakers of Bletchley developed the fundamentals of today’s analytics
You may or may not know that Station10 as a company was inspired by the code-breakers of World War II. A place described as one of the world’s best kept secrets – where mathematicians, technicians, translators and puzzle fanatics all came together and turned unintelligible data into highly valuable insight, thus shortening the war by at least two years – Bletchley Park was the home of the original data scientists.
At Station10 we like to organise a variety of team outings and activities; previous events include Escape Rooms and code-breaking cocktail making at The Bletchley Experience, both of which we (naturally) aced. Now, I am by no means a history buff – in fact it was my least favourite subject at school, most likely because there was no problem solving or numbers involved to play to my strengths, or possibly because at 14 years old I found it difficult to relate to crop rotations and the plague. Nevertheless it seemed fitting to arrange a more educational trip to the very place that inspired the creation of Station10.
Early on in the tour you pass wooden clipboards detailing a six step code-breaking process. What surprised me when reading them was how easily I could relate this 70-year-old process of intercepting and unscrambling enemy messages to what I do in my day to day role here at Station10, uncovering insights from customer data using modern day analytics..
Step 1 – Intercept your enemies’ radio signals
Eavesdrop on messages going to and from enemy headquarters and army, navy and air forces in the field.
(In light of the GDPR, eavesdrop is probably one of the worst wording choices to use, so I’ll go for the classic ‘listen’ instead). I probably don’t need to tell you that it’s important to listen to your customer base, not only on your direct site via website analytics, but across all channels – whether that be email, social or direct contact such as through a call centre. In order to effectively monitor your multichannel data, you should have proper tracking and a good implementation in place.
Step 2 – Work out how messages have been encrypted
Look carefully at the messages and use brainpower and intuition, together with the new machines you’ve invented, to break the day’s ciphers.
Step 3 – Decipher
Once the cipher has been broken, set teams of people to decipher all the messages received that day.
As an analyst, for me steps 2 and 3 blend into one. Once your tracking is set up correctly and the data being collected is coming through as expected, ask yourself ‘What is the most appropriate way to delve into all of this data to obtain useful insights?’ The answer is most likely to differ depending on the task at hand. Take your website analytics data – you may want a dashboard with top line traffic, conversion and revenue stats that senior stakeholders can look at daily. Your strategy teams may also need to use the same data to perform a bigger piece of analysis using a backlog of data from a range of sources, in which case you may need to use other more powerful software such as SQL or Tableau. It’s up to your analysts to use their own intuition and knowledge of different software to appropriately decipher your multichannel data.
Step 5 – Cross reference
See if anything in the message relates to other facts you know, to help build a better picture of what your enemy is doing.
Customer data in isolation doesn’t always explain behaviour. As this step says, in order to build a better picture, take knowledge from other sources before coming to conclusions and scheduling any changes. Let me use a basic example; as a website analyst I might see that online orders are down significantly vs the previous month, but what I might not know is that there has been an in-store only promotion on a range of high priced items in selected regional stores. Or, I could assume that an increase in Contact Us Form submits means an increase in potential new customers; however survey data in fact reveals that existing customers are having difficulty getting in touch via telephone.
The eagle-eyed amongst you will notice I moved step 5 before step 4, and for good reason. By cross referencing data before passing the message on, the likelihood is you’re reducing back and forth by pre-empting follow-on questions. Not only this but by joining data from multiple sources, such as offline sales data or survey data, you will be able to better understand customer behaviour and consequently the insights you draw off the back of this are likely to be more valuable.
Step 4 – Translate
Use language specialists to turn the deciphered messages into English so that intelligence experts can assess how important they are.
Step 6 – Send on the top secret intelligence you’ve uncovered
Make sure the information gets to the right people. Which senior military commanders will it help? Is it something the Prime Minister needs to know?
Similarly to previous steps of this code-breaking process, in my opinion steps 4 and 6 overlap. Consider these questions.. Who needs to be made aware of the insights that have been uncovered? It could be the sales and marketing teams, or the technical support teams, or both. With this in mind, what is the best way for different members of the business to consume the data relevant to them? It isn’t always the job of senior staff to understand the specifics of the data, so it’s important that analysts are able to communicate findings in an effective succinct way.
Whilst we might not be helping solve a World War, here at Station10 we use our experience and knowledge of technology to help you better understand and improve your multichannel customer data. If you need help breaking ciphers or building an enigma machine, or if you’re interested in joining the team, please get in touch!