How does ‘Customer Insight’ work in practical terms? Here at KETL we help our clients see the value of their data. Everything begins with good data storage and processing. Once you have the basics right then you can report from that data confidently. Data volumes constantly expand and so do the underlying data sources. So, your system must be able to grow, both in terms of data volumes and in the ease with which new sources can be ingested into the data store. The more accurate data sources you can use the more informative your insight engine can be. Our commercial clients are trying to get to know their customers and understand their behaviours and then use that understanding to influence or impact future customer behaviour. We build the infrastructure on which the data is housed that can enable the gathering of the right data, at the right time to then be used by the right people in a company to make the decisions needed. The customer insight engines or single customer view platforms that we build can then start to analyse past customer behaviour to infer future customer action, for customers who present similar characteristics. This video explains how we implemented Project Customer for the retailer Monsoon Accessorize https://vimeo.com/206258808 Customer data can also be mapped alongside external data sources, such as weather or demographic indicators. Social media and website information can also be used to infer customer intent or customer sentiment that can then predict behaviour, based on existing customer behaviour models. And the list goes on! I think there can sometimes be some confusion surrounding AI and Machine Learning. This article in Forbes is quite useful for helping to explain the difference between the two. Often, we find it easier for clients to start with the what and the why and then it is our job at KETL to work out the how. So, for example, if you want to start to think about who is visiting your shop or restaurant we can look at voucher redemptions and use in-house WiFi data to infer a customer ID match. If you want to do some competitor analysis we can build a query to go and search for keyword indicators from Google and identify similar stores/restaurants in a given area. This information could be used for gap analysis to decide whether a brand has enough coverage in a given area. Individual store/restaurant market conditions could be recorded against sales figures. This could then be enriched with external data to help with stock control and planning how to resource the outlet to account for local events. We could use key word or text analysis of journals and newspapers headlines and then merge that information with local and national data sets on sporting fixtures, events and other local festivals. For Send a Cow (SAC) we helped the researchers to start to put discipline around the evidence that they were gathering in their surveys and helped them to develop repeatable questions that could go on to be used by the charity’s analysts and researcher to look for patterns. The Impact Reports that SAC was able to create could then be used to help inform their donors of how their money was being spent, but it also meant that the analysis of the efficacy of the projects themselves could be more informative. https://www.sendacow.org/our-impact One of the ways that our industry has changed over the last few years is that the technology has significantly improved and become easier to configure. We can use products like Exasol that help to process enormous data volumes in a way that wasn’t possible 10 years ago. See this Exasol blog for an example of how large data sets can now be used http://www.exasol.com/en/blog/2017-05-02-the-power-of-data-to-raise-awareness-of-mental-health/ We are software agnostic. We don’t build software ourselves we use market-leading products to build the right environment for the client. Many of the products that we use to create an ecosystem for data storage, processing and reporting / analytics do have free to use versions. The software companies do this to enable prototypes or test environments to be constructed relatively cheaply. A sample stack could looks something like: Data storage: Exasol (free small business edition) Data orchestration: Talend (community version) Data analytics and visualisation: MicroStrategy (free) So it really is easier than ever before to think about a proof of concept (PoC) project to see what can be achieved. Just call us today and find out what you could do next.