Improving Marketing ROI: 5 Lessons To Learn From Airbnb & Uber

Performance marketing Sep 7, 2021

“The companies that are going to win are the ones who are using data, not guessing”

As organizations are going through the transformation of becoming digital-first, they are generating a large amount of data. Generating insights from such large scale & dispersed data is getting more difficult, thereby making it difficult to generate positive performance marketing ROI.

A very popular example of this is Uber ad fraud of 100 million USD, where Uber spent more than 100 Million USD on ad spends - only to discover that when they switched off their ad spend, they couldn’t see any drop in conversions. This is not just happening with Uber, almost all the organizations are burning their performance marketing spends and not getting the desired results.

Uber Ad Fraud
Tweet by Nandini Jammi

Performance marketing which is a boon for these internet-first companies is getting more difficult day by day with increasing spends. There are many conversations in the advertising fraternity that a large part of performance marketing is being wasted. A similar trend was seen when we saw Airbnb suspend their marketing spends amounting to 800 million USD after rise in covid cases.

These glaring examples forced us to reflect on the lessons for growth marketers when it comes to Performance Marketing.

#1 Don’t rely on Dashboards & Reports

As a growth marketer, if you are solely relying on Dashboards provided by your ad platforms, then it leaves a lot to be desired. It’s important for marketers to have all data from multiple ad platforms in ONE SINGLE place. Having modern analytics tools that offer pre-built integrations with platforms like Google Ads, Facebook Ads, LinkedIn Ads, Amazon is essential on Day 0.

Pre-built dashboards are generic and measure a handful of metrics - making them useful in a limited manner. At today’s scale and pace, you need tools that are capable of 24*7 monitoring and real-time alerting. Constantly refreshing dashboards to keep a tab on your campaign performance is not the 21st century.

#2 Go deeper than top-line KPI monitoring

As we noticed, tracking only a few top KPIs like Total Ad Spend, Impressions, CPC will hide more than it reveals. A sub-dimensional analysis is table-stakes to really prevent overspending & underspending on your campaigns. While an overall ad campaign might be performing at 3% CTR, there are definitely sub-target segments where it’s performing at 10% or 0.5%. It’s important to identify high-performing sub-groups and double-down on those while kicking out low-performers in real-time before it’s too late and thousands of $$ have been spent rather than wasted. Companies like Amazon, use advanced analytics like Anomaly & Outlier detection to spot trends and insights at a sub-dimensional level to make such decisions.

Real-time anomaly detection at a sub-dimensional data
Graph: Anomaly & Outlier Detection

#3 Perform Root Cause Analysis, correctly

As we noted in the Uber & Airbnb example, millions of $$ could have been saved if the root cause analysis for growth was done correctly and attributed to the right factors. Using data to monitor performance early on, could have prevented the guesswork - where the traffic was blindly attributed to ad spend based on experience rather than hard data, analysis & experimentation.

#4 Break analytics silos

A lot of the time, growth is wrongly attributed to marketing spend which leads to an incorrect marketing ROI. This is because organizations work in silos and there is no way to collate all the activities and events happening in different parts of the organization. For example - an ad campaign can increase traffic on the website - but result in decreased sales - due to poor site performance driven by high traffic. This requires co-monitoring & correlation of unusual spikes/drops across multiple related KPIs. Measuring ROAS is one small step in the right direction, but there’s a long road ahead to create a complete picture of internal & external events to understand the real root cause.

#5 Apply Machine Learning for analyzing large-scale data

It’s clear that with the scale of data that moderate marketing activity generates, the use of Machine Learning is imperative. In an ideal world, you’d be able to analyze every customer journey and the performance of all your marketing channels and website pages 24/7. This would mean that you’d be able to see exactly where issues are arising and address them before they have an impact on business metrics. But this isn’t a feasible or scalable solution when you’re manually tracking and analyzing KPIs. Humans aren’t made to crunch numbers, day in, day out, and we’re always going to make mistakes and miss out on important things.

Machine Learning, however, is perfect for this task. It’s trained to identify trends and patterns in data that are invisible to the human eye, so it can pick up on issues in your marketing data in real-time and alert you to prevent overspending/underspending on campaigns and convert your $ to real growth.

Performance Marketing w/o autonomous analytics like Anomaly Detection & automated Root Cause Analysis is a perfect recipe for an ever-increasing spend with limited upside and nothing in terms of marketing ROI. Rely on outdated analytics tools & dashboards at your own peril.

Tags

Great! You've successfully subscribed.
Great! Next, complete checkout for full access.
Welcome back! You've successfully signed in.
Success! Your account is fully activated, you now have access to all content.