This week, Faraday is thrilled to announce a series of product updates that mark a revolutionary moment in our own history, as well as the broader AI practice. These updates don’t just make our users’ lives easier, though that’s certainly a major part of it. They also bring us one step closer to the world that all of us at Faraday are invested in building: a world where brands can use AI in a manner that’s both widely accessible and intrinsically responsible.

What does accessible, responsible AI look like? How did we get here? And why does it matter? Read on to find answers—and join us in the revolution.

## Accessible AI for all

AI becomes accessible when _anyone_ can use it to make a prediction. You don’t need a million dollar marketing budget to get your foot in the door, you don’t need a data science degree to build a model, and you don’t need to write a line of code. Today, for the first time, Faraday users can cast all those barriers aside and make predictions about their customers’ behavior in just a few clicks.

This level of accessibility has been four years in the making, and while it seems like a no-brainer today, it was a risk when we committed to the new product roadmap. Our 2018 decision to invest in a sweeping, accessible AI product update was a gamble—and maybe a somewhat audacious decision at the time.

Back then, we were growing increasingly aware that brands wanted more control over their predictions. But pivoting from what was an intensely collaborative experience to a fully self-serve platform would require thousands of engineering hours, all for a technology that was still more of a buzzword than a common business solution. The software had to take a quantum leap forward.

Four years later, [AI is everywhere](https://hbr.org/2021/09/ai-adoption-skyrocketed-over-the-last-18-months) and Faraday’s quantum leap is here. We’re thrilled to be able to offer AI to the entire brand landscape, from retail to finance and beyond, paving the way for more and more companies to connect with their customers.

## Responsible AI for all

There’s a reason they say that with great power comes great responsibility, and AI is no exception.

Machine learning algorithms train on historical data, using patterns from the past to make predictions about the future. It should come as no surprise that, left unchecked, these algorithms will surface injustices that have been baked into decades of consumer data, reflecting common biases around race, gender, class, religion, and more. It’s completely possible to mitigate these biases and produce responsible predictions that avoid or even correct for past injustices, but it takes expert know-how.

For the first six years of our company’s life, we were tightly involved in every predictive model and we could be stewards of the crucial ethics involved. With the world moving toward AI for all, we had to make sure that it was truly AI for all and deeply embed responsible AI principles into everything we were doing.

As a result, Faraday has doubled down on responsible AI. We’ve always employed proven approaches for mitigating bias and stood by our ethical consumer data practices, but now we’re rolling out automatic bias detection and mitigation for our users as well. We’re serious about building a future where AI helps—not harms—our community, and we wouldn’t release a new wave of features without ensuring that our users are equipped to use them responsibly.

## The revolution is here

When Robbie, Seamus, and I started Faraday, AI was largely reserved for the Amazons and Walmarts of the world. The cost to develop, train, and supervise predictive models was exorbitant, and only the biggest corporations could afford that kind of investment. They’ve been reaping the benefits—and the profits—that come with predicting customer behavior ever since.

This summer, we’re ringing in an AI revolution at Faraday, one where any brand, not just the Fortune 100, can use practical, powerful AI without compromising on ethics. We’ll continue building on these updates for the rest of 2022 and beyond, and we hope you’ll join us for the ride.

This summer, we’re ringing in an AI revolution at Faraday, one where any brand, not just the Fortune 100, can use practical, powerful AI without compromising on ethics.

Faraday launches accessible, responsible, powerful AI for all

With summer in full swing, Faraday’s product team and developers have been working hard to create some incredibly exciting updates to the Faraday platform. The _why_ behind these changes is rooted in two of our core product pillars:

1. **AI for all**. The benefits of machine learning should not be reserved solely for the Fortune 500 or those with advanced degrees. The promise of easy AI, scalable cloud-based infrastructure, and a good UI should unlock the benefits of consumer behavior predictions for all brands.
2. **Making data science fun**. Using Faraday should be intuitive and easy. You should feel empowered to ideate, experiment, and deploy, all on your own.

We're thrilled to announce that as of today, four major updates are live. In this blog, we'll walk you through how each one works—here's what we'll cover:

- **Build your own predictions in Outcomes**: Describe the business outcomes you want predictions for, and Faraday will build and maintain predictive models for that outcome, with models getting more accurate over time as they learn from your data.
- **Deploy your predictions with ease in Pipelines**: Choose your likely-to-buy outcomes, your personas, and more; pick a specific population of customers; and target a destination for your deployment, like your data warehouse, cloud bucket, CRM, ESP, etc.
- **Plug into Faraday with Connections**: Plugging in your data is now a simpler process with fewer steps required. Plus, Connections paves the way for a future update where we'll implement true bi-directionality for your data and predictions in the Faraday platform.
- **Create more powerful Cohorts**: Cohort building now features the ability to specify certain trait requirements from your data or the Faraday Identity Graph in order for a customer to enter a cohort, adding an extra layer of depth to your predictions.

Let’s dive in.

## Build your own predictions in Outcomes

Back in December, we released an update that enabled Faraday users to build their own predictive personas within a cohort of their choosing based on various attributes from the [Faraday Identity Graph](/features/data).

That update had been the first major in-platform step toward user autonomy, and, more importantly, toward making AI accessible for all. Continuing on that path, Faraday users can now access an additional self-serve interface, this one for defining propensity (likelihood) predictions: **Outcomes**.

Outcomes allows you to easily describe the business outcomes you’re trying to predict in just a few clicks. Likely to buy? Sure. Likely to churn? You bet. Likely to buy _again_? That, too. Behind the scenes, Faraday takes over from here, applying dozens of strategies to build a range of candidate models, then selecting the one that most accurately predicts your outcome. Over time, Faraday automatically builds and rebuilds progressively more accurate models as you accumulate more data.

Our new Outcomes interface puts the power of experimentation in your hands, allowing you to easily try out different outcomes and view how effective the model will be across your use cases—all without needing a data scientist at your shoulder.

## Deploy your predictions with ease in Pipelines

Once you’ve defined your own prediction objectives, make them actionable with **Pipelines**.

Pipelines completely streamlines the prediction deployment process in Faraday. Previously, you may have used some combination of Reach, Monitor, and Inform to deploy predictions on-demand, on a scheduled basis, and in real-time via API. These three products have been unified into Pipelines for a more accessible and intuitive user experience.

Jump into Pipelines and choose your:

- **Payload:** a combination of your desired propensity outcomes, your personas, and more.
- **Population:** the people you want the predictions on.
- **Targets:** where you want your predictions deployed, including your data warehouse, cloud bucket, or anywhere else in your stack (CRM, ESP, etc).

Faraday takes over from here, keeping your predictions constantly up to date by refreshing scores and persona assignments when new data comes in, ensuring you’re never left wrong-footed when delivering personalized experiences to your most likely-to-buy customers.

With Pipelines, for the first time ever, you’re able to both describe and deploy your own predictions completely autonomously—without a data science degree.

## Plug into Faraday with Connections

Starting today, you’ll notice the Sources and Destinations areas of the Faraday platform have been combined into a single **Connections** view.

With Connections, you’re able to plug Faraday right into your data warehouse or cloud bucket by entering your location-specific details. Once done, it’ll be ready for selection as a deployment target in Pipelines. Dropping the number of setting screens required to complete your end-to-end connection from two to one should save you some clicks, and get your first predictions deployed that much faster.

Keep an eye out for more news here, because Connections is a major building block for a future update that will enable every Connection to be bi-directional—your predictions, wherever you want them, whenever you want them.

## Create more powerful Cohorts

Previously, **Cohorts** in Faraday were defined by event streams, or an event that each person in a cohort has experienced. This included the type of event (such as an order), when the person experienced the event (all orders, last X days, etc), the frequency of the event (such as number of times an order was placed), and finally, the value of the order.

We introduced the Segments console last year to allow users to define a group of people based solely on the Faraday Identity Graph (FIG)-based traits, like geography or income. While this was a great step for offering flexibility in terms of the types of population you could define in the Faraday platform, it was cumbersome having to navigate through both Cohorts and Segments to do so.

With today’s update, Segments and its functionality has been folded into Cohorts. Now, in one console, users can define their cohorts by selecting specific traits from either their customer data, or exclusively from FIG. Event streams are no longer required to build a cohort.

Prior to today, your "Recent bulk purchasers" cohort might’ve been simply based on the number of orders in the last 30 days. You can now refine this cohort further to say "Recent bulk purchasers in CA" or "Recent bulk purchasers with income >\$100k." On the flip side, you might also build a cohort without an event stream, such as "Dog owners in CA with an income >\$100k."

We hope that by allowing you to define your preferred population of people—whether that be based on an event stream like orders, FIG traits, or a combination of the two—gives you another building block upon which you can easily experiment with predictions through Outcomes, Personas, and Pipelines.

## What's next?

Coming up, we’ll be revamping data ingress to make it easier to map your data into Faraday. We see this as a critical step towards democratizing data science and empower you to create and maintain a powerful predictive layer atop your martech stack.

---

\*Explore our [product documentation](/docs/) or [talk to sales](/talk-to-sales) to learn more.

Introducing four major updates that empower users to build and deploy their own predictions in the Faraday platform—no data science degree required.

August 2022 product updates: Outcomes, Pipelines, Connections, Cohorts

Since the dawn of commerce, sellers have invested significant time and effort into figuring out what influences prospective buyers. Identifying your ideal customer—who they are, what they do, why they do it, and how they feel about various products and services—can help you craft powerful, people-first marketing campaigns that accelerate every sales cycle.

Building a customer insight strategy is the most efficient way to understand that ideal customer, evaluate your brand’s current situation, and adapt for the future. Here’s how you can get started.

## How to develop a customer insight strategy

To develop a customer insight strategy, you will need to focus on both qualitative and quantitative data. Gathering this information ensures you have a complete picture of who your ideal customer is, how they behave, and what their wants and needs are.

Identifying the key team players involved in the process and employing the right tools, techniques, and resources can take a lot of the guesswork out of building an insight strategy.

## Why customer insights are important for business growth

Competitive advantage begins with clear customer insights. Without a comprehensive customer insight strategy, brands are bound to miss out on crucial data that has the power to significantly impact their bottom line.

A customer insight strategy can unlock a wealth of information about your customers and what influences their purchases. Through a combination of consumer data, feedback, social listening, and website analytics, you can begin to draw conclusions about how your customers behave and why. This knowledge can help you improve the customer journey and create an experience that is truly memorable for those who purchase your product or service.

Having a strong foundational awareness of customer insights is the first step to developing a strategy of your own.

## Key components of a customer insight strategy

### Objectives

A clear understanding of why you want to gather customer insights can make developing a strategy easier. What is your reasoning behind gathering insights, and what do you hope to do with this data once you have it?

Maybe you’re hoping to increase revenue by a certain percentage next quarter or determine whether your current pricing strategy aligns with your competitors’. Maybe your ad campaigns aren’t converting like they used to and need to be refreshed. Or perhaps you’re exploring ways to optimize your current content strategy and attract a new audience. All these objectives can be achieved through a smart customer insight strategy.

### Research and data

A viable customer insight strategy calls for a combination of high-quality research and data. From customer interviews and social listening to surveys and focus groups, gathering data about your customers’ preferences can help influence your marketing strategy and content creation.

This emphasis on first-party data is increasingly important as the [end of third-party cookies](https://faraday.ai/blog/death-of-third-party-cookies-strategies) approaches, but you may also consider [responsibly sourced data](https://faraday.ai/features/data) from other vendors to add even more depth to your customer profiles.

Of course, high-quality data is of little use if there is no one to analyze it. A strong analytics team can help brands make sense of the noise and pull focus to the trends that matter most.

### Audience

There is no one-size-fits-all approach to customer insights. Before you can begin gathering valuable data, you’ll need to take a close look at which segment of customers you want to explore. For example, analyzing your current customer base will require a different approach than analyzing potential customers. Knowing whose insight is valuable in achieving certain objectives will ensure you use your time and resources as effectively as possible.

Keep in mind: customer feedback doesn’t always have to be about you to be useful. It doesn’t have to be gathered exclusively from your own customers, either. Finding out what people think about your competitors, for example, can provide a glimpse into what drives customers within your industry to buy.

### Customer journey map

Data-driven insights have the power to affect every level of decision-making when creating and marketing a product. As you gather data about your target audience, it should help you build a better understanding of your brand’s customer journey. Leverage those insights to create or refine a map of customer interactions with your brand—that way, you can skip the guesswork when it comes to what messaging and offers your audience should get at every step of their journey.

For instance, if most customers hear about your brand through digital channels but make their final purchase in-store, your campaigns should facilitate that leap from online to offline interactions. On the other hand, if most customers interact with multiple digital touchpoints before buying online, you’ll want to use those insights to guide them from touchpoint to touchpoint.

### Customer personas

Once you’ve thoroughly collected and analyzed your customer data, and determined what the typical customer journey looks like, you can take your customer insight strategy one step further to develop customer personas.

Personas are a means of clustering your customer base into distinct groups that share certain similarities. One persona group may contain customers who mostly have young families, live in the suburbs, and like to shop at brick-and-mortar stores; another may be digitally-native young professionals or wealthy urbanites. By honing in on the key customer insights that define each persona, you can tailor your campaigns to their unique interests and [better personalize every customer interaction](https://faraday.ai/predict/category/persona_assignments).

### AI enhancements

Sifting through such an enormous amount of data and unifying insights across platforms can be a huge endeavor. For some brands, off-the-shelf AI tools are the most straightforward way to enrich and make sense of all their customer data. Some of the ways they achieve this with AI include:

- Enhancing customer data with an identity graph to create detailed customer profiles
- Scoring leads based on how likely they are to purchase a product
- Predicting the lifetime value of potential or current customers
- Developing powerful quantitative persona groups
- Building lookalike audiences of good fits for use on any marketing channel
- Identifying which customers are at highest risk of churning

## Conclusion

Just about every successful brand relies on customer insights to inform their decision-making. Once you’ve pulled together all the components of your customer insight strategy, you can review them holistically to determine which insights offer the most potential for your team to work with.

An effective customer insight strategy can take the pressure off your sales and marketing teams, improve overall customer satisfaction, drive customer retention, and increase profitability over time. There’s no question that it’s worth your team’s time to get it right.

At Faraday, we offer a number of customer insight tools that allow brands to enhance their first-party data with responsibly sourced, opt-in consumer data and develop powerful, personalized marketing campaigns.

Learn what goes into a successful customer insight strategy and how you can start gathering the data you need to truly understand your customers.

How to create a customer insight strategy

Since 2009, [Momentum Solar](https://momentumsolar.com/) has been one of the fastest growing solar installers in the country, helping thousands of homeowners make the transition to solar.

## The problem: optimizing the lead scoring process

Momentum Solar relies on their state-of-the-art call center as their main conversion channel, reaching out to leads acquired from aggregators to set up on-site assessments for solar installations.

But Momentum didn't become one of the nation's largest installers by being satisfied with _good enough_.

A solar installation is a significant investment, and it can be easy to get cold feet. When a sales rep needs to dedicate an afternoon to an on-site visit that can be canceled at any time, the cost to acquire a new customer can get expensive quickly. To help offset those costs, the Momentum team was looking for areas of optimization with the goal of reducing acquisition costs. One area they looked at for improvement was a lift in their lead conversion rate. They believed that [better lead prioritization](/solutions/lead-management) that allowed their call center to focus on the best leads was key to driving this lift.

## The obstacle: an incomplete picture of a solar customer

Traditionally, Momentum had leveraged a rules-based approach to lead scoring that focused on a few data points, such as lead source, lead demographics, home ownership status, and location of residence, to determine lead priority. The challenge with a rules-based approach is these data points are often subjectively chosen, and don't account for the full spectrum of both intuitive and unintuitive data points that are key indicators of whether or not a lead will convert.

Additionally, Momentum had a large collection of leads that had not initially converted. A major issue with the rules-based approach to lead scoring is that it's static, even though people aren't. With Momentum's existing approach, a lead's score–and in turn, priority–would never change.

## Why Faraday?

### Fresh data + models

To optimize their lead conversion, Momentum used the Faraday platform to add a [predictive layer](/features/customer-behavior-predictions) to their lead prioritization system.

The value of predictive models is the quantitative, rather than qualitative, approach to lead scoring. Machine learning algorithms run millions of simulations to find patterns in large datasets, meaning they transcend subjective biases.

Instead of scoring a lead based on a few key points, Momentum used the Faraday platform to build a [lead scoring model](/blog/why-you-need-predictive-lead-scoring-today) that ranked leads based on their probability of becoming a customer. The model trained on Momentum's existing customer base and the [hundreds of first and third-party data points](/features/faraday-identity-graph) that Faraday associated with Momentum's customers.

Furthermore, the lead scores were updated as new customers' examples were added and lead behavior was observed. This ensured that Momentum's prioritization methodology was never static.

## The results

### Appointments set 33% faster

To test the efficacy of this new approach to prioritization Momentum split the scored leads into three groups: red, yellow, and green, with red being low probability leads and green being high probability leads. The goal was to see a _higher_ rate of lead-to-scheduled appointments rates for the yellow and green groups (comprising the top four deciles  of leads) and _lower_ conversion in the red group (bottom six deciles).

Each of these groups received the same call cadence and performance was measured against a baseline performance of 180 connected calls per scheduled appointment.

As expected, the green and yellow groups converted at increased rates compared to the baseline vs the red group, which converted at a lower rate. Specifically, the green group saw a **+191% lift in conversion rate** over baseline, while the yellow group saw a **+51% lift in conversion rate** over baseline. Equally as important, the red group converted 4% _lower_ than the baseline, further confirming where the call center should spend its time.

Ultimately, the call center needed **33% fewer outbound calls** to hit their scheduled appointment goal.

Momentum Solar used Faraday to add a predictive layer to their call center's lead scoring process, resulting in the call center needing 33% fewer outbound calls to achieve their appointment goal.

How Momentum Solar cut calls-per-appointment by 33% with predictive lead scoring

Third-party cookies and digital marketing have long gone hand in hand. For years, brands have relied on these small pieces of code to help them develop buyer personas and ad targeting strategies. But with data privacy growing increasingly important and consumers pushing back hard against data sharing, marketers must now navigate a world in which third-party cookies are quickly becoming obsolete.

The shift away from third-party cookies has its merits. Data privacy is a real issue for today’s consumers as concern over identity theft, viruses, and other cybersecurity threats put a growing number of internet-connected devices at risk.

A cookieless world can even be good for marketers too. It opens the door for brands to interact with their audience in a more human-centric manner, ultimately paving the way for better personalization. But the clock is ticking on third-party cookies and brands that haven't adapted already will need to pivot quickly in order to prepare.

## What are third-party cookies?

Third-party cookies are small bits of code that track internet users’ searches, website visits, and purchases, among other data. Whereas first-party cookies are dropped onto a website by its owner in order to monitor activity on their own asset, third-party cookies pass information back to an outside organization, like an advertiser.

Third-party cookies heavily influenced ad targeting for nearly three decades. Then, amidst growing concerns over data privacy, it all came crashing down.

In late 2019, Firefox announced that it would[ block third-party tracking by default](https://blog.mozilla.org/en/products/firefox/todays-firefox-blocks-third-party-tracking-cookies-and-cryptomining-by-default/). Apple followed suit in early 2020, though its [Safari browser has included cookie-blocking features since 2017](https://www.zdnet.com/article/apple-blocks-third-party-cookies-in-safari/). Additionally, Apple introduced [new privacy features in iOS 14.5](https://developer.apple.com/app-store/user-privacy-and-data-use/), allowing users to customize data privacy settings for individual apps.

Also in early 2020, Google announced it would be [phasing out cookies from its Chrome browser](https://blog.chromium.org/2020/01/building-more-private-web-path-towards.html)—a move that will[ take effect in 2023](https://www.bbc.com/news/technology-57611701). This shift away from third-party data accessibility has led marketers to contemplate the future of ad campaigns.

## Why does the death of third-party cookies matter?

With user data no longer as readily up for grabs, brands must adapt their approaches to ad targeting and predicting buyer patterns and behavior. Marketers will need to be more strategic with their data and explore new ways to segment audiences.

Unfortunately, not all marketers are prepared for the task. According to a survey by Epsilon, [54% of marketers](https://www.adweek.com/resource-library/marketers-still-arent-ready-for-a-world-without-cookies/) still don't feel fully prepared for the death of third-party cookies.

Brands that hesitate or refuse to think outside the box for future data collection will likely fall behind the crowd. Ensuring cost-effective advertising in a changing digital world means being vigilant about new technology and data sources. The key is to connect users across multiple, respectful touch points, such as opt-in data opportunities and contextual ads.

## Strategies to prepare

### Take stock of your first-party data

How effective have you been at collecting first-party data? Identifying methods to collect opt-in first-party data, such on your website and through email marketing, will help you respect your users’ privacy while still meeting marketing goals. If you’re currently using a content management system (CMS), the analytics dashboards should provide valuable insight into key metrics that can work to your advantage and help you identify where you may need to fill in some gaps.

### Get consent

Want access to solid user data? Just ask. By providing incentives, such as discounts or a loyalty program, you can invite users to share additional data with you and make it a win-win situation—they get a tangible benefit, you get opt-in data that’s fair game for marketing outreach. Just make sure you’re transparent about how you’ll use it. Building trust with your audience is essential to growing brand loyalty.

### Embrace AI

Predictive modeling uses artificial intelligence to better understand your audience and uncover prospective buyers without relying on cookie-based data. AI tools can work backward from your desired outcome to identify which consumers are likely to be your next best customers, based on a combination of your first-party data and responsibly sourced opt-in data.

### Implement contextual advertising

Contextual advertising is when ads are matched to the content on a given site page, as opposed to the individual site visitor. Because ads are related to the content a user is already consuming, many consider this type of targeting less intrusive than behavioral targeting. It offers promise for marketers who want to continue serving the most relevant ads possible while still respecting consumer privacy.

### Provide value

Third-party data made it easy for marketers to rest on their laurels. Now, brands must try even harder to remain relevant and earn their audience’s attention. There’s no better way to achieve this than to continue to invest in quality products, services, and content. At the end of the day, nothing is more important than providing value to your customers.

### Explore new strategies

Paying close attention to innovative tools and data sources can help marketers keep up with ever-evolving privacy challenges. Fortunately, a growing array of resources are helping marketers adapt their first-party data strategies and make ad targeting more effective. Remain open to exploring different options and you’ll likely find one that aligns with your goals.

## Let Faraday guide you through a cookieless future

The death of the third-party cookie isn’t all bad news for marketers. Taking the time to make more meaningful connections with your audience and investing in future-proof solutions will mean wiser ad spending in the long run.

The end of third-party cookies is coming. Here's what you can do to stay connected with the right audiences.

6 strategies to prepare for the death of third-party cookies

[Advia Credit Union](https://www.adviacu.org/)'s mission is to provide financial advantages to its members through solutions that address life's evolving needs, including mortgage refinancing, debt consolidation, and auto loans.

## The problem: engaging customers with the right message

The average American holds eight financial products, yet only [19% of Americans](https://www2.deloitte.com/content/dam/Deloitte/us/Documents/financial-services/us-kickingitupanotch-092614.pdf) have three or more products with their primary financial institution. For many credit unions, including Advia, growing their "wallet share" with [existing members](/solutions/customer-engagement) is key to their long term stability and success. The challenge is how to do this effectively. There is brand risk in simply bombarding members with unsuitable offers that could result in members ignoring all communications from the credit union. Worse, it could erode trust, create frustration, and sow overall member dissatisfaction.

For Advia, they needed a way to understand the personalized financial journeys for each of their members so that they could surface the [next best product recommendation](/features/customer-behavior-predictions) for each member. Such an approach would ensure optimal use of their marketing budget, minimize member dissatisfaction, and help grow wallet share with their members.

## The obstacle: using data and remaining compliant

Advia understood that their member data likely held the key for developing next best product recommendations. Unfortunately, their lead data scientist had recently left the organization, and Advia was looking at a lengthy ramp period to hire and onboard someone new.

Additionally, Advia needed to navigate a complex regulatory landscape. Advertising lending products comes with stringent compliance requirements to ensure fair lending practices and protect consumer privacy. The Equal Credit Opportunity Act (ECOA), for example, prohibits discrimination in credit transactions based on race, color, religion, national origin, sex, marital status, age, or receipt of public assistance. This means any data used for segmentation in advertising must not incorporate these protected classes, either intentionally or unintentionally, to avoid violations.

## Why Faraday?

To address these challenges, Advia partnered with Faraday. Advia used the Faraday platform to determine which of their primary retail financial products should be surfaced to each member. They initially tested this by looking at which members should receive information on their auto loan product.

Advia linked its historical member data with Faraday's identity graph, including information on member adoption of Advia's auto loan. With this training data, Advia created a set of predictive models that scored each member's propensity for an auto loan. They also leveraged Faraday's [bias identification and mitigation features](/blog/faraday-bias-management-announcement) to address historical bias in their training set, aiding their fair lending compliance efforts.

With Faraday's product recommendations in hand, Advia could segment their audience more intelligently. They identified the top percentile of members most likely to complete an auto loan application and targeted them through both personalized email and direct mail campaigns. This targeted approach ensured that the most relevant offers reached the right members, enhancing engagement and increasing the likelihood of loan applications.

## The results

The success of Advia's auto loan campaign was measured by the percentage of recipients that completed auto loan applications. Before using Faraday, Advia's benchmark for success was a 1.19% application rate.

**30-Day results**: After just 30 days of running the campaign that used Faraday's recommendation, Advia achieved a 2.41% application rate. This significant improvement resulted in **$1.1 million in new auto loans**.

**90-Day results**: The application rate continued to grow over the next 60 days achieving a 5.18% application rate after 90 days. This increase in engagement led to a total of **$2.7 million in new auto loans**, far surpassing the initial benchmark.

Advia Credit Union's use of Faraday's platform demonstrates the power of personalization and predictive analytics. Faraday gave Advia the ability to target the right members with relevant offers, leading to higher engagement and significantly improved loan application rates.

This approach not only boosted Advia's campaign performance, but also boosted Advia's regulatory compliance efforts, while safeguarding member trust. Advia's use of the Faraday platform exemplifies how financial institutions can harness data to drive better outcomes, provide superior service to their members, and use [AI responsibly](/features/ai-safety-and-responsible-ai).

The team at Advia Credit Union used customer predictions to deliver the right product to the right people in a direct mail campaign while using AI responsibly.

How Advia Credit Union generates $2.7 million in new auto loans with Faraday’s consumer product recommendations

In the competitive landscape of direct-to-consumer (DTC) ecommerce, [personalization](/blog/personalization-at-scale-guide) is a key differentiator.

[Boll & Branch](https://www.bollandbranch.com/), a prominent DTC linens company, used Faraday to provide customer predictions that guide their marketing strategies.

## The problem

Boll & Branch was looking for a more accurate way to effectively segment their customer base so they could deliver more personalized and relevant content.

With a wide range of products and a diverse customer demographic, it was crucial to understand more about their customers, their preferred products, and channels.  They found that their initial segments had no correlation to specific products or marketing channels, limiting their ability to launch an effective personalization strategy.

## The solution

To address this challenge, Boll & Branch used Faraday's persona predictions to identify customer personas that represented distinct socio-economic clusters. Each persona had preferred products and marketing channels.

This approach differed from a traditional persona exercise in which a brand must either form personas from a qualitative process (e.g. surveys, focus groups, etc.) or leveraging pre-computed personas like PRIZM codes.

With [qualitatively derived personas](/blog/predictive-personalization-machine-learning), a brand can't easily assign new customers to a persona, unless a customer responds to survey questions. Given most customer surveys have single digit response rates, this limits a brand's ability to segment their customer base.

Pre-computed personas have the benefit of being universal, but are generic demographic clusters vs. specific patterns derived from a brand's data. This often results in persona segments with little variation in customer behavior.

Faraday's persona models leverage [rich data](/features/faraday-identity-graph) on nearly every U.S. adult–allowing for universal assignment–paired with an AI clustering algorithm that builds personas from a training set provided by a client's own customer data. The results are customer personas that can be applied to all known contacts in a brand's customer base, while also reflecting the socio-demographic nuances across this dataset.

### Data integration and persona development

Boll & Branch used the Faraday platform to identify three distinct personas based on a set of socio-demographic factors. They then were able to see that these segments had distinct purchasing behavior, preferences, and other relevant metrics.

These personas were then assigned to individual customers in Boll & Branch's [BigQuery](/integrations/predictions-with-bigquery) data warehouse, providing the metadata needed to support their personalization initiatives.

### Personas + attribution

Traditionally, Boll & Branch thought their demographic skewed older, and as a result they invested heavily in radio advertising. Pairing Faraday's personas with Boll & Branch's internal attribution system allowed them to see what kinds of customers were coming from which channels, what discount codes different personas were responding to, and what products they were purchasing.

An example of this was the marketing team's identification of their highest value persona, whom the team nicknamed "Jen": a wealthier, settled, homeowner, that preferred clean modern aesthetics.

This qualitative insight gave Boll & Branch the confidence to organize marketing towards the Jen persona - from branding, product design, prospecting, all the way down to engagement and retention.

As Daphne Chow, Senior Manager of Consumer Strategy & Analytics puts it, Faraday helped Boll & Branch "see what kind of consumer we are seeing as crucial to our success, and our whole organization has rallied around that goal."

For example, they could see that individuals in the Jen persona were heavy users of social media, and as a result began being more strategic with their channel mix.

Knowing Jen also allowed them to conduct focus groups and surveys with members of that cohort, where they found they were influenced by interior design personalities - guiding them to a successful partnership with the star designer, Nate Berkus.

### Personas + email campaigns

Because every contact in their system was assigned a persona, Boll & Branch could also use these assignments to power their email campaigns. The team designed personalized content for each persona: not only did they use creative that aligned to the socio-demographic makeup of each persona, they also were able to draw insights about what products each of their customer personas over-indexed on, allowing them to select highly relevant product recommendations for each persona.

This approach ensured that each customer received content that resonates with their unique preferences and needs.

## The results

The implementation of Faraday's personas into Boll & Branch's marketing strategy yielded impressive results. By delivering personalized experiences through targeted email campaigns, the company achieved a **30% lift in conversion rates**. This significant improvement underscored the value of utilizing advanced predictive analytics to drive personalized marketing efforts.

Boll & Branch's partnership with Faraday exemplifies the power of data-driven personalization in ecommerce. By identifying distinct customer personas and tailoring marketing messages accordingly, the company was able to enhance customer engagement and drive higher conversion rates.

Or, in the words of their Chief Digital Officer, Katia Unlu, "Faraday is in our DNA."

Boll & Branch used customer predictions to drive hyper-personalized customer experiences, resulting in a 30% lift in coversion rates.

How Boll & Branch achieved a 30% lift in conversion rates with customer predictions

A household name to many, Cabot Creamery provides high-quality dairy products to the New
England market. But despite their limited regional presence, they have amassed a loyal
following email subscribers — roughly 300,000 strong — from across the country.
To better reach these brand enthusiasts, Cabot worked with Faraday to learn more about
their subscribers, as well as gauge new market opportunities outside of New England.

To better reach their brand enthusiasts, Cabot worked with Faraday to learn more about their
email subscribers. Offline, Cabot is looking to expand their physical presence in grocery stores and other retailers beyond their current New England market. To identify the best cities for success, they leveraged Faraday’s market sizing capabilities.

## Gaining a deeper understanding of Cabot's most loyal fans

To grow their email subscription list and engage customers, Cabot captures emails from
online promotions, in-store sampling, and giveaways—but their real-world insight into these
subscribers was limited. Enhancing their first-party data with rich third-party attributes, Faraday gave Cabot a holistic picture of their email subscribers, going so far as to make comparisons between subscribers in different regions of the country to see what sets each geo apart.

Key insights included age, an interest in cooking and mail-order food, and — surprisingly —
pet ownership. Many of these common interests among email subscribers varied from place
to places across the country. With this in mind, Cabot can confidently personalize future
content to resonate with specific segments of their email subscribers.

## Identifying high-potential markets with location-based predictions

To engage new customers outside of New England, Cabot needed to find the best locations
for future expansion. Faraday analyzed the brand’s areas of interest, predicting which zip
codes in New York City and ten metro areas in the Southeast were potential goldmines.
These geospatial predictions will not only inform Cabot’s market expansion, but can guide
location-based marketing initiatives, like SEM and traditional media buys. Even the placement
of in-store sampling events can benefit from these predictions.

Learn how the beloved brand used consumer predictions to guide its market expansion strategy.

How Cabot Creamery uncovered high-potential markets for expansion

[Community First Credit Union](https://www.communityfirstfl.org/) (CFCU) started in 1935 in a Jacksonville high school basement and has since evolved into 19 branches and over 135,000 members. CFCU's success is rooted in its commitment to its members, offering tailored products to meet individual needs and utilizing innovative technologies to deliver an exceptional banking experience.

## The problem: finding the right members

A mere 19% of Americans have three or more products with their primary financial institution–that is, where their checking account lives–despite the average American holding [eight financial products](https://www2.deloitte.com/content/dam/Deloitte/us/Documents/financial-services/us-kickingitupanotch-092614.pdf).

In order to remain competitive & ensure long-term success, CFCU knew that they had to grow their "wallet share" with [existing members](/solutions/customer-engagement). More members adopting CFCU's comprehensive financial solutions deepens the financial relationship between the credit union and its members, fostering loyalty and long-term engagement. This diversification helps in spreading risk and maintaining a more stable financial base, while reducing the need for costly marketing and acquisition efforts.

The challenge they faced was how to accomplish this effectively. CFCU lacked the resources to compete directly with larger banks and fintech companies. That meant simply shoveling more ad spend into their [direct mail](/blog/personalized-direct-mail-strategy) program wasn't an option - CFCU needed a more efficient approach that could increase response rates from both new and existing members _without_ increasing spending.

## The obstacle: reaching the right members efficiently

Before Faraday, CFCU used traditional segmentation, like a member's current product mix, their income, and identifying members to whom they could extend pre-approved offers of credit. What they found is that this approach often failed to capture the nuanced preferences and [behaviors](/features/customer-behavior-predictions) of individual members. Because segments were built off broad categories, the generic messages tailored to these segments didn't resonate with recipients. [Personalized](/blog/personalization-at-scale-guide) and relevant content were essential to better find and engage loyal members.

CFCU understood that AI offered opportunities for more effective segmentation, on top of an unlock for personalizing a member's journey.  Their existing member data provided the historical examples of customers who had grown wallet share from one product to many. Their existing direct mail providers had exposed CFCU to the immense trove of additional third party data that could be leveraged to create distinct segments.

At the same time, they knew that were signficant costs associated with this data, the infrastructure needed to run AI models, and the in house talent needed to develop that level of modeling.

## Why Faraday? 

### Identity resolution & easy product recommendations

Partnering with Faraday turned out to be the unlock that CFCU needed to realize this vision of using AI to grow wallet share amongst their members.

They used the Faraday platform to connect their member data to Faraday's identity graph, giving them instant access to [hundreds of data points](/features/faraday-identity-graph) beyond those they had used to segment members in the past. Using Faraday's modeling engine, CFCU was able to build product recommendations that were personalized for each member and for all prospects in their charter territory.

With this segmentation in hand, CFCU used Faraday to easily build mailing lists of the individuals they most wanted to target in their direct mail campaigns.

## The results

### Acquisition campaign

CFCU used the Faraday platform to predict which consumers were most likely to become their next loyal members, defined as a member with a high account balance and strong net interest revenue. To test the model, CFCU ran a direct mail campaign with a 50/50 audience split (50% from high-scoring model predictions; 50% from traditional segmentation techniques).

The new members generated using Faraday's targeting saw **61% higher average account balances** and **24% higher net interest revenue** than the new members acquired using traditional segmentation techniques.

### Next product recommendation

Faraday also predicted which members were most likely to need a checking account, auto loan, first mortgage, and more. To test the model's predictions and get in front of members with the highest potential for engagement, CFCU ran a direct mail campaign with the same technique in a 50/50 audience split (50% based on Faraday's next product recommendation; 50% from traditional segmentation techniques).

The members contacted with offers recommended by Faraday saw **54% higher net interest revenue** and **88% lift in projected ROI**.

With Faraday, Community First Credit Union used customer predictions to drive highly effective acquisition and engagement campaigns with great results.

How Community First Credit Union unlocks 61% higher account balances using Faraday’s predictive AI

It should come as no surprise that digital marketing is ever-changing. In times like these, when social media reigns supreme and trends are rapidly evolving, it makes sense that good marketers constantly seek ways to evolve in terms of tools, techniques, and strategy. Although digital marketing trends rarely remain stagnant, there is one particular tool that continues to be popular across industries: the lookalike audience.

Whether you’re brand new to marketing or have been in the game for years, there are plenty of options for you to explore around lookalikes. New marketers may be wondering what a lookalike audience is and how it can help them reach a previously untapped audience; seasoned pros might feel like they’ve exhausted their options and need a new way to move the needle with their lookalikes.

Luckily, today’s ad platforms have made it easier than ever to mimic audiences and target them efficiently and effectively. And there are even more solutions, like predictive audiences built with enhanced first-party data and AI, that can take campaigns to a whole new level. Let's dive into the world of Facebook lookalikes in particular, since Facebook remains one of the most commonly used platforms in the digital advertising space.

## What is a Facebook lookalike audience?

We'll start with the basics—if this is old news to you, feel free to scroll down for more sophisticated tactics beyond traditional lookalikes.

[Lookalike audiences](https://www.facebook.com/business/help/164749007013531?id=401668390442328) allow Facebook marketers to target users that have similar interests to those of their current customer base. If you’ve had significant success advertising to a core group of current customers on Facebook’s ad platform, you may now be eager to find more users that share demographics or behaviors with those found in this core group. This is where lookalike audiences come in.

There are a variety of components you can use to build a lookalike audience, from similar demographics—such as age, location, and interests—to certain psychographics that influence buyer attitude. Lookalikes based on these criteria can prove extremely valuable when it comes to an ad campaign’s success because you already have a good sense of what's most likely to appeal to these potential new customers.

## How to create a Facebook lookalike audience

Creating a lookalike audience is a fairly simple process so long as you have access to the right tools. A Facebook ads account and administrator permissions, for example, are necessary to get started. You will also need to determine whether you need to install a Facebook pixel on your website. The pixel is a simple way to identify audiences that are interested enough in your product or service to visit your site.

To set up a lookalike audience on Facebook, follow these steps:

1. Head to your Facebook Ads account and click on **Audiences**
2. Find the **Create Audience** dropdown menu and select **Lookalike Audience**
3. Select a seed list of users, such as a Custom Audience built from 1,000 of your best customers
4. Select the country or countries you’d like to target
5. Choose the size of your audience
6. Select **Create Audience**

## Benefits of Facebook lookalike audiences

There are several benefits associated with implementing lookalike audiences in your digital marketing campaigns, which is why this remains such a popular method for reaching new audiences on Facebook and other platforms as well.

### Tracking

Optimized campaigns call for analytics and tracking. Lookalike audiences on Facebook make this process quite easy. With lookalike audiences, you can monitor progress and cost, evaluate the road to conversion, and adjust your campaign as needed.

### ROI

One of the most frustrating parts about advertising is not really knowing whether your investment will pay off. Sending advertisements to the wrong audience costs time, money, and frustration. Creating lookalike audiences is one way to increase the likelihood that real prospects will see your ad, all without wasting resources on users who may have very little interest in your product or service.

Keep in mind that lookalike audiences are particularly useful if you are already working with a strong customer base, since you can upload a list of customers that you know are valuable in order for Facebook to surface more people like them.

### Customizations

Lookalike audiences allow for impressive customizations that make targeting audiences more efficient than ever. This is because customizations allow you to adjust everything from goals to budget and even the percent of the population that you target with your lookalike. On that latter point, our recommendation is to keep the percentage low. This allows Facebook’s algorithm to hone in on a narrow group of people that most closely resemble your seed audience, which can [produce higher cost-person-conversion and clickthrough rates](https://adespresso.com/blog/adespresso-experiment-facebook-lookalike-audience/).

### Enhancements

To take it a step further, you might even consider using a lookalike audience builder to gain access to additional customizations beyond that of the Facebook platform. At Faraday, those customizations include enhancing your first-party data with [responsibly sourced consumer data](https://faraday.ai/features/data) to help you more effectively understand and segment your audience, as well as building custom lookalike audiences that can be used beyond Facebook.

## Using Facebook lookalike audiences effectively

Relevant, effective ads cost less than those that don’t connect with your audience, but you don’t want to spend time and resources on creating lookalikes just because you can. Incorporating lookalike audiences into your ad campaign must be strategic to elevate your product or service.

To that effect, you'll still need to hone your copy, creative, and CTA to determine what resonates most with your audience. Since lookalike audiences are generally used to expand your brand reach and identify potential new customers, they work best for top-of-funnel campaigns. Consider A/B testing your ads to better understand this new audience and keep moving them through the buyer's journey.

## Alternatives to Facebook lookalike audiences

There’s no denying that lookalike audiences are a powerful tool for brands. However, there are a few key caveats marketers should be aware of, and additional solutions that may work even better depending on your circumstances.

First, it’s important to keep in mind that Facebook is able to build powerful lookalikes because it identifies patterns in user data that it collects from third-party cookies. With cookies due to be phased out over the next few years, Facebook’s lookalike capabilities are likely to grow far more limited.

Additionally, Facebook’s lookalike audiences work only on the Facebook platform, making it difficult for marketers to craft strategic campaigns that target an audience across channels. Given that the average buyer’s journey contains [multiple touch points](https://www.thinkwithgoogle.com/intl/en-cee/consumer-insights/consumer-journey/study-reveals-complexity-modern-consumer-paths-purchase-and-how-brands-can-make-inroads/), that’s a significant hurdle to overcome in an increasingly noisy digital world where marketers are competing for limited ad slots.

Luckily, there are other options for teams that need a more sophisticated means of orchestrating the buyer’s journey. Predictive audiences are a particularly promising method, leveraging responsibly sourced consumer data and machine learning algorithms to craft highly accurate lookalikes that can be used across channels.

## Level up your lookalike audiences today

Whether you’re just starting to experiment with lookalike audiences or are ready to expand into predictive solutions, nothing beats diving in and testing out your audience options.

For Facebook’s lookalike audiences, you may want to test different segments as seed audiences for your lookalikes to determine which list works best.

Whether you’re brand new to marketing or have been in the game for years, there are plenty of options for you to explore around lookalikes.

All posts