DataMa Compare introduction
- 1. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent DataMa Compare Introduction for consulting usage
- 2. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent What is DataMa Compare? DataMa Compare is a software licensed and developed in R by DataMa Company, designed to help you understand why A is different than B on a given KPI. Once installed on your machine or within your organization, it consumes your client’s data to provide a simple visualization that will make your life easier. You can try it free of charge on any client’s use case by contacting us on: www.datama.org
- 3. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent Use Case 1: Week on Week website performance monitoring Client’s input DataMa Compare output This client, an e-commerce website, wants to understand what’s driving the gap of Revenue/ User from one week to another (read full use case description on DataMa Compare demo article) An access to the Google Analytics API providing e-commerce metrics, split by key ecommerce dimensions “ Revenue/ User drops by -19% mainly due to conversion issues, caused by SEO traffic changes”
- 4. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent Use Case 2: Actual Margin vs. Budget Client’s input DataMa Compare output This client, a consumer good producer, wants to understand why his current absolute margin is higher than initially budgeted. An access to IBM Watson financial forecast, with volumes, revenue and margin split by product and countries “ +27% higher margin vs. budget thanks to much higher volume in Germany and higher price point of product B than expected”
- 5. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent Use Case 3: Insite Page Content benchmarking Client’s input DataMa Compare output This client, a media online player, wants to understand why Page A is performing better than Page B in terms of conversion, and which specific section of the page is causing the gap An access to Content Square data, splitting performance metrics by page zones and devices “ Page A is doing 2 times better than page B thanks to higher engagement on text zone, and higher views/ visits, particularly on desktop”
- 6. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent Use Case 4: Revenue comparison between two geographies Client’s input DataMa Compare output This client, a large hotel player, wants to understand what’s driving the difference in revenue between US regions and EU region An Excel (CSV) extract, splitting business key indicators by region, brand and Area type “ We have less hotels in the US, and they are smaller in number of rooms and room size, with slightly lower occupancy rates.”
- 7. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent More broadly, when should I use DataMa Compare? Start (e.g. LY) End (e.g. TY) ? How can I quickly explain this… … with this ? Dim 1 (e.g. Medium) Dim 2 (e.g. Device) Start End Dimensions MetricsAny KPI (e.g. $/Session) A (Session) B (PV) C (Bkg) D ($)
- 8. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent DataMa Compare: A plug & play R model explain KPI gaps easily FOR DISCUSSION Why? What? • Compare any KPI between two segments and understand quickly what's driving the gap, with large data sets • Mix effects • Segments performance changes on specific dimensions • Underlying metrics changes • Typical use cases (see next pages): • Understand the performance (Revenue, Conversion, Traffic) variations from one period to another (Week on Week, Year on Year, Budget vs. actual...) • Compare two business units, or two products (e.g. two web pages) to understand how to improve one of them • Inputs: a flat table with metrics distributed on dimensions (typically, a CSV) • Inside: an R code. R is open source, free and easy to install software on your current system • Outputs: an other flat table with drivers identified. This can be plugged on your current visualization tool (e.g. see Tableau template) or use R Shiny apps (see demo app here) • Any executive in charge of explaining regularly variations in a KPI, and identify drivers in a complex flow. Typically sales or marketing executives, e-commerce directors, BI heads...Who?
- 9. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent DataMa Compare approach: 3 main levels of analysis Waterfall analysis Dimension analysis Mix effect Segment Performance Summary • Once KPI equation properly written, identify which underlying KPI is responsible for the change • For each underlying KPI, analyze how much mix effects on each dimension contributes to the observed gap • For each underlying KPI and each dimension, identify which specific segment performance is driving the observed gap • For each underlying KPI, select most interesting dimension to explain observed gap
- 10. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent DataMa Compare key advantages for consulting ✓ Big Data: R can easily tackle large Data Sets. Low limitations in number of metrics & dimensions ✓ Client ready: Charts can be easily inserted in slides ✓ No black box: Constant access to calculations and row data, hosted by your own machine ✓ Accessible: Even without knowing R, you can easily set up your own use case ✓ Business oriented: simple maths, easy to explain, for smart business decisions ✓ Flexible: With Shiny interface, you can easily drill down/ filter/ simplify the output ✓ Quick: Use cases can be set up in <5 minutes ✓ Reproductible: You can implement an automated dashboard to update analysis regularly
- 11. This information is confidential and was prepared by DataMa solely for the use of our client; it is not to be relied on by any 3rd party without DataMa's prior written consent References: DataMa website DataMa Compare : • GA use case • Shiny live demo • Tableau demo Contact: guillaume@datama.org