Online performance naturally starts us thinking about data. A lot of talks over the two days at A4U London were orientated around data, the term ‘big data’ features prominently in an online marketers repertoire of late, but not many have taken the time to ask what it all really means? Data and digital marketing go together like bread and butter or peas in a pod, it makes perfect sense for their harmonious relationship after all, data is one of the things that separates traditional and digital marketing. In many respects affiliate or performance marketing simply wouldn’t exist if we weren’t able to collect data to assess affiliation or performance of online campaigns.
The panel for this discussion was occupied with a number of data junkies and self-confessed geeks that were primed to help us try and understand it all, including:
- Panelist: Dominic Edmunds, Founder & President for SaleCycle
- Panelist: Mike Glegg, VP Sales for MyThings
- Panelist: Giedre Guntulionyte, Customer Acquistion Manager for Reiss
- Panelist: Laurie Walton, Account Director for TagMan
The Importance of Data
Big data seems to be going mainstream but when asked, the panel treated it very much as an overused buzz word, stating “it’s more of a buzz word than a fact, data is data. Big data is a creation of a term for something that we’re already familiar with”. However they went on to conclude that big data should be described as “a large amount of data that can be pulled together to create a meaningful response or specific actions”. According to the panel, data only officially becomes ‘big’ when it is 1 petabyte or greater and that’s a serious amount of data! In order to put this in perspective, BBCiPlayer reportedly requires 7 petabytes of bandwidth per month; so most claims of ‘big data’ are in fact, false.
Data can be understood by the 3 V’s
- Volume – The amount of data being considered
- Variety – Of data, such as: multi-structured data or different formats of data that when pulled together creates actionable insights
- Velocity – Speed, or ongoing speed, of how that data is collected, utilised and/or processed
Where to Start when Handling Data
Most people often ask ‘where do we start?’ when it comes to data, so the panel chose to explore some top tips on this. Primarily they agreed on starting small, trying to understand absolutely everything from the word go will make you feel like the world is closing in around you.
Reiss, who were represented on the panel, explained that they started by first looking at all of their digital channels: By identifying the channels with the lowest ROI first, they were able to prioritise and work from there. What was important, and effective, with this approach was that they had identified that they wished to use data to simply to solve problems rather than trying to figure out some fancy way of using it, which would provide no actionable outcome. For your 3 steps to data success, consider this approach:
- Think about what you’re trying to achieve
- Identify the desired outputs
- Work backwards and assess what infrastructure is required to analyse this
Nowadays the level of detail of data that is collected is vast and overwhelming, while most people consider looking at data historically, some ask whether there is a place for real-time reporting in this? Generally the panel came to the conclusion that real-time was a ‘nice to have’, but you cannot forget about historical data altogether…
- Using real-time data you can optimise for some channels but often real-time can become misleading
- Historical data provides the ability to create forecasts and predict trends. This can be used strategically to understand the bigger picture and make quantifiable predictions and insights
- There is still a place for both historical data analysis and real-time tracking/analysis
Utilising your data
“Ensure there is transparency in the data and there is a strong ROI, otherwise we may as well all pack up and go home…” (Dominic Edmunds, SaleCycle).
- Consider using your data to recognise your existing customers and then identify opportunities for new customers based on this
- By creating prospecting display runs can enable you to understand what area of a site, individual section or page is performing best – these can then be optimised for and targeted specifically
- Using data you can begin to build up dynamic product recommendations, such as analysing users baskets to provide similar products or dual/multi-selling, consider Amazon who pioneered this and continue to utilise this to good effect
Understand that there are risks with using data (such as the Sony PlayStation 3 information leak example). The risks are minimal providing you do your homework. Consider devising a data strategy to allow you to:
- Understand matching data sets
- Visualised your data
- Recognise opportunity for cost savings, this can then be reinvested to pay for the technology that you invested in in the first place
- “Air New Zealand stated at JUMP Conference last week that they said that they have increased revenue by 20% by taking their savings and reinvesting it in themselves.”
- If you’re struggling to get to grips with your data, consider booking a few days with a data analyst freelancer, or hiring an in-house resource
- Take the time out for your busy day to step-back and really look at your data
- Understand that data will become increasingly more layered, use this to your advantage!
- There is still opportunity to be had in proving offline media spend, this matters for online marketers because this is an avenue that we can exploit to free up higher budgets for digital spend.
- “Data is the new oil” – IBM announced that every day we create 2.5 quintillion bytes of data, this is so much that 90% of the world’s data today has been created in just the last 2 years!
- For all people starting out looking to be an analyst, it’s not about fancy tools – just pull a load of data and stick it in an Excel sheet, roll up your sleeves and get your hands dirty.