Analytics Idol 2017 In Review
Though Adobe Summit 2017 has come to an end, 12,000 digital marketers made the journey to Las Vegas this year and left more knowledgeable, connected, prepared — and eager to push their companies to become the experience businesses of the future. I am proud to have led the Analytics Idol competition — a breakout session with close to 600 attendees who want to learn some tips and tricks to take back to their organizations and vote for them in true reality-show fashion.
View our slide-show presentation — “Analytics Idol 2017: Top tips & tricks” — to listen to this year’s competition and learn more about the contestants and their tips. Below, you will find a review of our five finalists and the world-class tips each revealed.
Brad Millett, Senior Analytics Strategist, Blast Analytics & Marketing
Tip 1: Use data from your Activity Map to automatically fill in gaps in data.
Brad recognized that customer data occasionally drops off the map. To fill in these gaps, he was able to leverage data that is automatically populated by the Activity Map! Brad then demonstrated how to use a combination of calculated metrics and segmentation to resolve the issue. This tip works very well because the data is captured simply by having a recent version of your Adobe Analytics code.
Tip 2: Define and visualize your customer’s journey.
At Blast A&M, Brad has discovered that customer journeys tend to vary, and every business uses different levers to help their customers along the many stages of their journeys. In this tip, Brad suggests a framework for identifying these stages, mapping them to key performance indicators (KPIs), and using Analysis Workspace to represent them visually. In addition to mapping each stage to KPIs, Brad highly recommends creating segments that can be applied easily for deeper analysis. Finally, it’s important to act on this data, which means empowering your teams to interact with the data via Workspace curation.
Hila Dahan, Cofounder and Principal Analyst, 33 Sticks
Tip 1: Though difficult, connecting the disparate dots along a customer’s journey is vital.
Hila showed us a strategy for linking activity across several different platforms to track a customer’s journey from end to end. First, Hila suggests including your offline data in Adobe Analytics by creating a handshake between your offline database and Adobe. Then, take advantage of Customer Attributes in Adobe Analytics to map your most important characteristics. Connecting this data allows for analysis and personalization and creates a customer journey that is more properly stitched together.
Tip 2: Democratize analytics data through an integration with Slack.
Hila and her team use Slack — a workplace communications tool that is enabling companies all over the globe to stay in touch. To empower your team to be better informed of data anomalies, Hila built a brand-new integration with Slack that she recommends using. Just visit the 33 Sticks – AskAdobe page to install the Slack application and enable it for Anomaly Detection. The benefit is that analysts receive alerts in a tool they’re already using so they can openly communicate with others. In addition, anomalies that are especially difficult to understand can be ‘starred’ for follow-up in Slack.
Rob Adams, Senior Digital Analyst, W.W. Grainger
Tip 1: Attribute product recommendations to actual 1:1 revenue.
Rob’s group is a heavy user of Adobe Target-driven product recommendations and focuses deeply on analyzing the success of these algorithmically presented products. To do this analysis, Rob and his team needed to directly attribute each product recommendation to actual 1:1 revenue. He was able to do this using a merchandising eVar (bound to ‘Add To Cart’), a classification file (to identify location and experience information), and ReportBuilder (to conduct extremely granular analysis). For even more details, watch the Analytics Idol slide show.
Tip 2: Break out month-over-month data more fairly.
How many times has your company looked at month-over-month data? Well, maybe it’s time to break out that data in a way that’s a little fairer. Rob suggests that companies take advantage of the time-parting functionality they already possess to also separate business days from non-business days. Doing so may help explain why a 2.4 percent drop in revenue in February was actually a 2.5 percent increase in revenue in comparison to January — all because there were more days to do business in that first month. A combination of your time-parting eVar, classifications rule builder, an Analytics sub-classification, and ReportBuilder all make this possible.
Prolet Miteva, Senior Manager of Web Analytics Platforms, Autodesk
Tip 1: Motivate your team to be passionate about data.
Prolet’s team noticed a problem — her marketers were not excited about data. To solve this, she decided to go where her users were and give them data there. Deciding to start small — focusing on basic traffic and scroll information, navigational sources, and firmographic and technological data — Prolet chose to build an in-house extension for Chrome that would bring the data right to their fingertips. Her marketers were then able to access to each of these data types at the click of a button — right in their browser! Prolet found that this drove adoption of data, better questions from her marketers, and further interest in Adobe Analytics deployment.
Tip 2: Know the potential audience size you can target for personalization and testing.
Understanding the potential audience size for a target offer is a requirement for choosing your segmentation wisely. Prolet has identified a slick way to do this — leverage Adobe Analytics segmentation and the integration of Adobe Analytics and Adobe Target. First, define your Adobe Marketing Cloud audience, evaluate the audience size, create a sequential segment to handle the transfer of audience data between systems, and then create a comparison workspace. The benefit is that you get the power of Analytics segmentation with the verification of your population size — all before you deploy your test.
David Bressler, Analytics Lead, Net Conversion
Tip 1: When event data is missing, use stats, page views, and correlation analysis to fill in gaps.
David knows that ‘things’ happen with data, and sometimes, you have to be creative to resolve those things. When you can, David suggests simply using a combination of page views and an additional calculated metric to fill in the gap. When this won’t suffice, David recommends using his process for leveraging a correlation analysis and linear equation to empower your calculated metric with even more accuracy.
Tip 2: Create a calculated metric that truly shows average page-load time.
David walked the Idol attendees through a valuable and arduous process to better understand how page-load time affects conversion. Based on his analysis, a one-second improvement in page-load time helped drive a 3 percent increase in conversion. David then revealed his secret sauce — a combination that includes an app measurement plugin, a classification rule builder, a few segments, and one MASSIVE calculated metric to obtain the data he needs. You really must see it to believe it!