Extending Google Analytics with API Integration

Preface

graph-imageThis is an introductory article in a series of posts aimed at outlining how Google Analytics can be extended using API Integration to suit complex marketing analysis needs that cannot be met within the existing Google Analytics interface. It outlines some key scenarios where extracting Google Analytics data offline remains the only viable alternative to not doing insightful analysis at all. The focus of these posts will very much be on outlining the business needs (as opposed to delving into technical solutioning) and identifying suitable scenarios where time and investment in building a data/technology savvy capability can be commercially justified. In subsequent posts, we will attempt to use examples of actual implementations which we hope will encourage Web Analytics practitioners in more actively exploring the art of possible when it comes to Google Analytics.

Extending Google Analytics with API Integration

Google Analytics has rapidly grown into being a poster boy for Enterprise Web Analytics offering a powerful set of basic tracking features that can be extended to support nearly limitless scenarios for cross-channel tracking and measurement. Google has been constantly refining its Analytics offering through introduction of compelling new features including Real-time Analytics, Multi-Channel Funnels and the most recent addition of Measurement Protocol. Yet, there are numerous use cases where critical business needs cannot be met using the native Google Analytics Interface. Rather than complicate their offering by adding features that many users may seldom use, Google has cleverly opted to open up access to its Analytics Engine so that complex, bespoke Analytics scenarios can be implemented offline through API integration.

In the sections below, we identify 4 key business scenarios where extracting Google Analytics data offline would make most sense from a cost/benefit perspective.

1.       Need for Advanced Reporting Features not supported within Google Analytics Interface

Despite its recent enhancement, the reporting capabilities within the native Google Interface remain fairly primitive and inadequate for advanced business intelligence delivery. The tool lacks advanced analysis features including multi-dimensional drill-down reporting, what-if analysis, forecasting, outlier detection, pattern recognition etc. It lacks any capabilities in the area of visualization. Numbers rarely tell a story and effective visualization is key to quickly understanding patterns and focusing optimization efforts for maximum returns. Lack of an easy way of reporting on consolidated data from across multiple profiles, lack of support for calculated metrics and on-the-fly dimensions all remain inhibiting features in adoption of native Google Analytics interface for even semi-advanced web analytics.

2.       Need to address tagging inconsistencies and clickstream data quality issues

Garbage in, garbage out is a fairly apt description of Analytics output. Tagging lies at the heart of any Web Analytics implementation and ensuring consistent tagging across multiple channels and campaigns remains an elusive reality for most businesses. Inconsistent tagging creates all sorts of data anomalies making it almost impossible to develop accurate Insights. An inbound link with utm_source=Bing will be treated differently to one using utm_source=bing within Google Analytics. A link using utm_campaign=XYZ is not the same as utm_campaign=DEF even though both links likely refer to the same campaign. It is physically impossible to QA a large campaign running across multiple channels and involving multiple geographies each with varying levels of language skills and technical competence. The easiest approach in such cases is to extract Google Analytics data offline and cleanse it as per bespoke data transformation rules before doing actual analysis.

3.       Need to track offline conversions and optimize based on post-conversion data

Optimizing Acquisition performance usually involves tweaking budget allocation to sources that introduce new visitors who then go on to generate better quality and quantity of conversions at lower cost. Getting this level of insight when those conversions happen offline remains a major challenge. With the introduction of Universal Analytics, Google has made it relatively straightforward to track Acquisition performance for Paid Advertising. However, things get a lot more complicated when it comes to tracking owned and earned media which cannot be tagged in the same way as paid campaigns. Another issue relates to tracking quality rather than quantity of conversions. Companies typically start by tracking conversion values and cost of acquisition up until the point of first conversion. This is certainly possible within Google Analytics but inevitably leads to overrating of sources that bring high volume of conversions, perhaps event at lower costs but where conversions tend to be largely one off and many a times for low margin products. A more sophisticated approach is to optimise for customer lifetime value which takes into account a number of other considerations over a period of customer lifetime. This level of analysis goes well beyond the capabilities of existing Google Analytics product and the only viable solution remains that of pulling data out of Google Analytics into a bespoke Analytics engine.

4.       Need to support advanced Data Mining and Predictive Analytics

Google Analytics is essentially a reporting tool providing basic reports to assess online performance. Many a times though, analysis requirements extend beyond reporting to custom data mining and these require access to raw clickstream data. While getting access to hit level data is still not possible (unless you are Google Analytics Premium customer), it is possible to setup automated, periodic extracts of aggregated data that can then be used for a number of use cases including Campaign Optimization (Media attribution), Econometric Modelling, Performance Benchmarking for Conversion Optimization, Click fraud detection, Creative optimization, and Behavioural profiling. All these requirements extend well beyond traditional Web Analytics but Google Analytics does a splendid job in providing access to underlying data and facilitating advanced Marketing optimizations.

These are the 4 major scenarios where pulling data out of Google Analytics for offline analysis remains the only commercially viable option. Companies looking to expand their Web Analytics capabilities beyond limited utility, basic clickstream tracking can pick and choose any of the above as a business background to defining an advanced Digital Analytics data strategy and gradually building out an Analytics Enterprise Architecture using Google Analytics as the primary data collection engine.

Building on the business needs summary presented above, the next few posts will present more in-depth discussion on each of the scenarios above starting with the next post where we will take some practical examples of how to use Google Analytics data to conduct advanced data analysis and reporting.

 

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>