big data is getting bigger. how is your organisation dealing with the insights available through big data?
5 tips on creating an analytics engine that provides valuable insights into your customers
The increased usage of multiple devices, smartphones, tablets, work and home computers to search and purchase products online is one of the major factors behind the increasing volume of data, along with retail checkout, credit card transactions, call centre logs and various other sources.
Companies are trying to make sense of the data, but the unfortunate truth is that some are just collecting a large amount of data. The data comes from different channels, through more than one system and everyone in the organisation is expected to both understand the insights and act upon them.
Extracting and acting upon relevant and commercial insights from the raw data collected throughout the customer journey, if not already, should become a priority for all companies in 2013.
This article provides 5 tips to help retailers and brands introduce a process that takes advantage of big data enabling it to be used effectively to determine which channel or budget was responsible for generating a particular action. A common mistake is to assume that one single action is necessarily responsible for a sale. Actions such as a click on a display ad or a ’like’ in a Facebook page can both be relevant and ultimately perceived as triggers for a sale. Beware these actions may also be part of a more complex set of cascading actions and triggers that lead to that very sale. Analytics platforms are becoming more and more sophisticated. Along with this increasing sophistication, the human capacity to analyse and make sense of data is also becoming more and more important. Retailers and brands need to embrace analysts in 2013 and not just software.
These 5 tips are not in any specific order. They are all important when introducing an analytics engine in your organisation to make sense of big data.
1 - A single customer view is more than the sum of various channels.
The customer journey has become more complex in recent years. Every time the user changes device during their path to conversion it gets more difficult to say which channel has effectively contributed to the sale. How can you know which channel was responsible for generating a sale when the customer received a direct mail leaflet, searched the company name, found the homepage but clicked on a sponsored link leading to a landing page, went to and liked a Facebook page, became a twitter follower and retweeted your last tweet but bought in one of your stores in London?
Companies commonly measure the performance of each of their activities as if they work independently of each other and this may result in a bad measurement because one action in one channel can exert an enormous influence, or assist the revenues in another. In Marketing those assisted effects can overestimate paid-search revenue and underestimate social media for instance.
2 - Integrate mobile and tablets into your multichannel offerings
Customers tend to carry their mobile phones with them at all times and so cross-channel calls to action, such as QR codes, can be used to engage with customers within store environments. Of all the trends I have been following in the multichannel industry, I believe that the Digital store is the next ’big’ step for brands and retailers, with mobile and tablets working as a connector between the online and offline worlds. Mobile and tablets are already facilitating services like click and collect, iPad concierges or interactive display solutions but we will witness further innovations, for instance with payment options, in the near future.
All the customer journey data generated through mobile devices can be measured across online and offline channels. When integrating these devices with your multichannel offering you are facilitating any attribution model you have in place.
3 - Embrace analytics with a C level executive sponsor
Senior level buying of analytics packages is essential to help promote clarity and alignment an early stage. The second step is to assign an analytics-minded director or manager to sit on a cross-functional team; giving guidance and allocating resources across units as the project expands. The intelligence, that is essential for the success of the analytics engine, often lies between marketing, finance and customer service. The challenge is to consolidate all the data coming from different sources and create systems that facilitate the on-going collection.
4 - Dealing with real-time data and non real-time data at the same time
Not all data is the same. When it comes to measuring data effectively, timing is relevant but not everything needs to be analysed in real-time. For example, qualitative data from forums does not emerge immediately. So in these cases it makes sense to establish a plan to look at these variables throughout different quarters. With non real-time data it is important to standardise analytics in order to forecast trends and patterns.
When it comes to measuring results in real-time, automation is paramount. Real-time data is especially important for advertisement investments; one minute bidding on the wrong keyword or placement could result in spending thousands of pounds ineffectively.
5 - Attribution is only the first step
New analytics models are emerging powered by the integration of big data and cloud computing, with attribution modelling as the one that brands and retailers are more focused on at the moment. Attribution is the process of quantifying the contribution of each channel for a specific action, e.g. click, sale, etc. The use of simplistic analytics will never encompass the correlated effects that go into a view that accurately explains customer behaviour.
Other aspects that are seldom considered are exogenous variables such as competitive offerings, seasonality, and geography. A certain complexity is required in the era of big data and once a company has quantified the relative contribution of each channel and the influence of exogenous factors, the next step is optimisation.
Optimisation uses predictive analytics tools to run scenarios for business planning. Using the actual elasticity’s of your business drivers you can run thousands of possible scenarios within seconds/minutes. Predictive scenarios allows companies to have a much clearer understanding of the strategic landscape and adjust all tactical plans quickly to gain a competitive advantage.
Finally, there is the activity of allocation. This is related with the real time redistribution of resources or budget across the organisation. When businesses start having multiple sales channels, multiple products, multiple brands or multiple geographies, analytics get considerably more complex and more than most internal teams can handle. However any company, regardless of its size, can start by building the right foundations of an effective analytics infrastructure coupled with an in-house adaptive marketing culture. The challenge is to adopt this not only quickly, but also more importantly, before your competitors do.
This article was previously published here http://www.neoworks.com/blog/post/20130916-big-data-is-getting-bigger