5 ways to prevent wasting e-commerce ad spend

First published on April 21, 2022


8 minute read

Felicia Kuan



Marketing budgets fly away in today’s competitive bid for ad space. Ensure every dollar is spent wisely by using ML to customize marketing campaigns and optimize your business’ return of ad spend.


  • Intro

  • 1. Personalized, timely ads

  • 2. Delivering dynamic offers for different parts of your funnel

  • 3. Reusing content by tailoring ads to every channel

  • 4. Leverage no-cost ads: email!

  • 5. Take advantage of your competitors 

  • Conclusion


After the pandemic necessitated e-commerce as an alternative to in-person shopping, digital advertising trends will continue to shape the retail landscape. However, as more marketing campaigns are finding success in the platforms’ personalized algorithms for showing ads to target customers, the ad costs are rising with the demand of ad bidding.

According to Clickbank’s survey, 50% of their e-commerce clients reported ad spend on popular social media platforms to be their highest expenditure in 2022. For context, one client of Clickbank mentioned that Meta charged them 

47% more

in December 2021 compared to the year before, totaling an average of $16.89 cost per impression. 

Therefore, how do we ensure efficient spending? On which channels should I spend more/less to optimize return on ad spend (ROAS)?

In this post, we’ll discuss 


common ad spend challenges in the e-commerce industry and how to address them.

1. Personalized, timely ads

Ever wonder how your social media feed seems to read your mind?

Data engineers designed 

OCEAN metrics

, a method of evaluating your personality based on how Open, Conscientious, Extroverted, Agreeable, and Neurotic you are. These metrics are used by social media companies to understand how to present ads to you in the most compelling way. Machine learning (ML) algorithms obtain users’ personality, things they’ve shared, time, and location to predict online activities. 

From personal experience, I’d definitely not engage with delicious food ads around my bedtime, or fashion and beauty products from companies that disregard sustainability or human rights practices. And I’m not alone; according to Salesforce’s State of the Connected Customer report in 2020, 


of customers expect companies to understand their unique needs and expectations.

Although we might not have all the data on customers, their social media and online purchasing behavior is scrutinized by Meta and Google’s AI-driven algorithms. While these platforms offer a straightforward advertising process, there are limitations to using Google and 

Facebook Business

platforms: You can’t access the raw data, or examine and tweak the machine learning engine to prioritize certain prospects. 

By using a 

third-party advertising platform

, you can draw correlations between effective marketing campaigns and account for the best times to show ads—this is particularly useful for companies with customers worldwide. Machine learning can help e-commerce businesses sift through a plethora of data to identify the region and 

optimal send times

for presenting advertisements. 

2. Retargeting users at different parts of your funnel

Our next tip is adapting your strategy to consumer familiarity with your brand to 

optimize your funnel

. Research by 


found that retargeting campaigns generated the highest lift in business name searches by 


. In fact, consumers are 


more likely to convert after clicking on a retargeting campaign ad.

Perhaps you have users who repeatedly clicked into your ad and browsed your website. Maybe they’re also one of the 

86% mobile

users who left their shopping cart abandoned. While it would traditionally take a herculean effort to identify and engage high-intent users, we can leverage machine learning to proactively retarget these warm prospects who already know about our product. We can use customizable AI-driven advertising platforms to configure search parameters for high-intent users. After using AI to identify those users, we can use 

predictive machine learning models

to automatically decide (with your input) how to re-engage these different prospects:

  • Calculate better offers for users in the decision stage that will get them to buy

  • Show them even more personalized ads

3. Reusing content by tailoring ads to every channel

With machine learning insights, you can skip sifting through the noise to get a clear persona of your users that informs marketing decisions across different channels. 

After using machine learning to gather customer insights, we can make inferences about the personas of potential customers on each of your channels. Explorium recommends asking these questions to 

identifying the right channels


  • How can we distinguish the kinds of users who are engaging with ads?

  • What platforms garner the most engagement, while also delivering the best-quality leads?

  • Where do you get the best ROI?

As advised in the book, 

In-bound Content

, you don’t need to create completely new content to share on different platforms. You just need to identify what kinds of people your ads are reaching on each platform, and tailor different messages using the same ad content across different channels. The author, Champion, suggests that you can reuse content by sharing only a snippet of an advertisement on a different platform that specifically targets a particular prospect.

This is what it means to wring every last drop of value from all the content you create!

4. Leverage no-cost ads: email!

While paid ads often yield results, you don’t want to exhaust your entire ad budget and not have any other marketing channels to fall back on. According to 

Instapage statistics

, personalized promotional mailings have 41% higher click rates and 29% higher unique open rates compared to mass sent, unpersonalized emails. Sending targeted emails to warm prospects can be a cheap and effective way to engage with potential customers in addition to regular social ads.

Machine learning also has a solution for this. There are 

AI tools

using natural language generation (NLG) to create personalized email drafts that learn your company’s brand voice. It uses deep learning to communicate in a way that resonates with your target audience. 

Lastly, Airship trained a model to learn and automatically send emails at an optimized send times to increase open rate. They found that the personalized send time performed at a 

52% higher match rate

compared to the general best time to send emails.

5. Take advantage of your competitors 

Using Google ads’ 

custom intent audiences

, we can gain access to your competitors’ first party data about their customers and prospects. You can even gather information about the keywords driving traffic to your competitor’s most popular pages for your own advertising campaign. As a small business, this is the best way of utilizing your limited resources: make the big guys do all the hard work.

Based on the market research of prospective customers from your competitors, you now have the freedom to be more creative in your own marketing efforts. We can use AI to predict what these prospective customers need, and proactively devise marketing campaigns based on these decisions.


We predict the e-commerce landscape will continue to grow. As more and more people conduct business online through mobile and social in the coming years, it’s crucial to stay on top of the rapid changes and understand the optimal ways to reach your customers– because they’re definitely out there, Google searching for you. 

Start building for free

No need for a credit card to get started.
Trying out Mage to build ranking models won’t cost a cent.

No need for a credit card to get started. Trying out Mage to build ranking models won’t cost a cent.