Sorbet

A textbook example of how we learn fast, apply what we learn, and continuously improve as performance indicators change.

SERVICES

Media Optimization
Data: Custom Events
Content

MARKET

North America

Industries

B2C Financial Services

CLIENT WEBSITE

www.getsorbet.com

Sorbet

A textbook example of how we learn fast, apply what we learn, and continuously improve as performance indicators change.

SERVICES

Media Optimization Data: Custom Events Content

Industries

Financial Services

Industries

B2C Financial Services

CLIENT WEBSITE

www.getsorbet.com

20 %
Decrease in Acquisition Cost
50 %
Increase in Converted Leads
20 %
Increase in Conversion Rate
20 %
Increase in Leads from New Creative Assets

Brief

Before engaging with Noetic, Sorbet was struggling to scale their initial success generating leads from paid media. To help them overcome this obstacle, we needed to optimize their ad performance in three key areas: media optimization, event conditions, and ad content variety.

Learning what changes were needed, making those changes, and tracking performance over time allowed Sorbet to make their digital advertising dollar go farther than it ever had before.

Strategy

01

Media Optimization

Effective paid media campaigns require the right audience targeting and segmentation, iterative testing, and compliance with ad platform standards.

02

Data: Custom Events

What qualifies as a conversion event? Is it a form fill? A completed application? How and when are these events fed back into the algorithm to optimize lead quality and quantity?

03

Content

Ad content needs to be dynamic, unique, and relevant to the viewer. The best performance comes from creative content that engages and intrigues.

01

Media Optimization

01

Media Optimization

In June 2023, Sorbet’s Meta ads account had conversion campaigns running with one objective: to receive new applications for Sorbet’s PTO cash payout service.
Although these campaigns were obtaining results, the final offline conversion rate was abysmal. This was a bottom-of-the-funnel problem, and optimizing media would be the first place to start in crafting a solution.

Challenge

Before working with us, Sorbet’s active campaigns were testing across fragmented groups in eight states, and learning which ads were performing better was being stifled because of an IOS14+ visibility issue.
We needed to give the machine learning algorithm bigger pools of data to yield better-performing ads over time.
Additionally, some ad sets were combining one campaign with another, which was generating murky testing results.
Lastly, Meta’s limitations with the ‘credit’ ad category was a barrier to finding the best-qualified users. We would need to get creative to solve all these problems in unison.

Method

To see the fastest results, we initially consolidated disparate ad campaigns into a single, unified campaign optimized to the same conversion event.
In parallel, we modified our testing structure to include a separate campaign strictly for testing purposes. This allowed us to avoid disrupting the consistency of evergreen campaigns while testing was taking place.
Then, to contend with the ‘credit’ ad category restrictions, we executed a creative targeting strategy using prepared messaging and visuals that clearly described life situations that would be perfect for a PTO liquidation.

Outcome

All in all, it took one week to begin seeing results and another 4-6 weeks for our media optimization strategy to be fully implemented.
As a result of consolidating campaigns and orienting them for the same optimization event:
Submitted applications increased from 4% to 53%.

Booked applications increased from 12% to 104%.

Approved applications increased from 69% to 182%.

To cap it all off, our creative approach to overcoming the credit ad category restrictions led us to our best performing ad in 2023, which has yielded a 43% Conversion Rate (CR) in 2023 and a 49% CR so far in 2024.

02

DATA: CUSTOM EVENTS

02

Data:
Custom events

With media now optimized, it was time to increase the quality of the users we were attracting and the amount of completed applications they were submitting, without increasing Customer Acquisition Cost (CAC). To crack this code, we would need to analyze event conditions and make Meta’s ad tracking technology work for us.

Challenge

In Sorbet’s case, both the event definitions and pixel configurations were set up in a way that was hurting both lead quality and ad optimization.
To make matters worse, Meta’s Conversions API (CAPI) wasn’t being used, which meant communication back to Meta was hamstrung.

Now that the problems were clearly identified, it was time to get to work fixing them.

Method

Our first task was to implement Meta’s Conversions API so the learning algorithm knew what ads were working well and what ads weren’t. This completed the feedback loop back to Meta, allowing the pixel to season and ad performance to improve over time.
Next, we restructured events so that only valuable users (those who met certain criteria, like income and employment type) could trigger the ‘bookable lead’ event. While this resulted in a reduction in the quantity of leads, the quality of each lead shot way up.
While we made these changes, visualizing our progress was possible thanks to a robust Business Intelligence (BI) platform that gave us deep, actionable insights.
Sorbet was informed of what was happening in real time, and soon, we began seeing the fruits of our labor.

Method

Outcome

We now had a higher-quality, better-targeted audience.
We also had a much more effective lead qualification process that was producing the kind of leads Sorbet wanted.

27 % boost of approved applications.

20% reduction in Client Acquisition Cost.

Today, Sorbet enjoys much smoother and more predictable lead generation using the Meta platform, and their Return on Ad Spend (ROAS) is healthier now than it has ever been before.

03

CONTENT

03

Content

Prior to working with us, Sorbet’s ad content was, in a word, lackluster.
Their existing stock imagery and generic messaging were providing meager results, and overall ad design wasn’t capturing the essence of the Sorbet brand.
We knew we could do better, and it would take a team effort to get there.

Challenge

Before working with us, Sorbet’s active campaigns were testing across fragmented groups in eight states, and learning which ads were performing better was being stifled because of an IOS14+ visibility issue.
We needed to give the machine learning algorithm bigger pools of data to yield better-performing ads over time.
Additionally, some ad sets were combining one campaign with another, which was generating murky testing results.
Lastly, Meta’s limitations with the ‘credit’ ad category was a barrier to finding the best-qualified users. We would need to get creative to solve all these problems in unison.

Method

To reach the demographics we were targeting, we decided to work with UGC creators who themselves fit our target demographic.
We looked for creators in specific age ranges and with certain ethnic backgrounds who had lifestyles that conformed with the Sorbet buyer persona.
We then tested the messaging to determine which words and phrases performed best. For example, we learned that ads using ‘Paid Time Off’ versus ‘PTO’ performed slightly better.
Some of the best-performing UGC content involved creators giving real-world examples of how they used Sorbet and how the service enhanced their lives.
With continued testing and tweaking, we started seeing significant results.

Outcome

Keeping what worked and improving what wasn’t was the key to unlocking success for Sorbet,
Through iterative testing, demographically aligned UGC content generation, and balancing the ad format mix, we were able to improve Sorbet’s return on their advertising dollar by 48%.
Additionally, the highest-converting UGC ad was responsible for producing 60% of Sorbet’s total conversions on the Meta platform.

Return on advertisement dollar increased to 48%.

UGC produced 60% of Sorbet’s total conversions.

To cap it all off, our creative approach to overcoming the credit ad category restrictions led us to our best performing ad in 2023, which has yielded a 43% Conversion Rate (CR) in 2023 and a 49% CR so far in 2024.

Project Summary

Sorbet’s paid media campaigns were restructured and optimized to reduce confusion and increase performance. Audience targeting was reconfigured, making campaign tracking more logical.
Event definitions and triggers were brought in-line with a more realistic model of the company’s sales funnel, giving Meta’s machine learning algorithm better data.
Event definitions and triggers were brought in-line with a more realistic model of the company’s sales funnel, giving Meta’s machine learning algorithm better data.

Technology Used

Case studies