Janhavi Nagarhalli

Content Marketer

Factors.ai

Hello to whoever is reading this. I'm Janhavi, a lawyer who went through a series of unfortunate events and realised that law isn’t for her. I stumbled on marketing by accident (as most of us do, I presume) and fell in love with it. From writing a simple 800-word blog post for a coaching center to creating content for SaaS companies around the world, my professional trajectory has been nothing short of a rollercoaster!

But that's about work. Aside from writing long-form articles, I like to travel, bake, read, and watch psychological thrillers. I spend my free time incessantly finding new things to learn about, be it philosophy, politics or new countries I can travel to for cheap.

To quote Walt Whitman, "I am large; I contain multitudes". If you’re curious to find out more or just want to yap about life, feel free to connect with me on LinkedIn and catch up!

Understanding Customer Churn Prediction in 2025

August 11, 2025

•

0 min read

Imagine working hard for months to close the deal with a prospect, only for them to churn in less than a year. There could be several reasons, such as:

Poor customer service
Choosing a competitor's solution
Users not achieving their KPIs

Reducing customer churn is vital for businesses because it ensures customer satisfaction and boosts profitability. The best way to avoid high churn rates is to predetermine the customers at a churn risk.

In this article, we'll detail how customer churn prediction is the key to reducing churn and keeping the cash flowing in 💸

What is Customer Churn Prediction?

Customer churn prediction involves analyzing data to detect customers likely to cancel subscriptions. SaaS businesses use this analysis to identify at-risk customers, leading to cost savings and improved customer lifetime value.

Analyzing churn through data-driven insights can help your business understand patterns and provide a roadmap for improvement. For example, if your surveys reveal that your platform has a complicated onboarding process – you can direct users to your onboarding specialist to assist them.

Why is Customer Churn Prediction important?

Losing customers is always costly. However, the costs involved go beyond the revenue lost from the customers who leave. It also includes the marketing expenses required to find new customers to replace the old ones. In many cases, the amount spent on acquiring a new customer is not covered by the amount paid during their time with the company. This means that customer acquisition is usually more expensive than customer retention.

Plus, unhappy customers share their experiences with others, impacting the company's reputation and customer acquisition budget. Businesses must predict churn and take action beforehand to prevent customers from leaving.

Once you know a customer is going to churn, you can take actions such as:

Providing more targeted re-engagement campaigns
Launching incentives such as loyalty programs that encourage them to stay
Creating educational material that is tailored toward their specific needs
Ensuring accessible and improved customer support

How to Build a Customer Churn Prediction Model

Creating a churn prediction model can help businesses retain customers and sustain growth. Using data analytics and machine learning, companies can identify which customers are likely to leave and take action to prevent it.

Here are the key steps to develop an effective churn prediction model ⬇️

Data collection and review

Ensure that the data is accurate by handling missing values, removing duplicates, and converting it into a suitable format for analysis. Before moving on to calculations, reviewing the data for accuracy and validity is crucial. Every piece of customer info is valuable in the upcoming churn calculations, so it's worth ensuring accuracy.

Model selection

Select an appropriate machine learning algorithm for churn prediction, such as logistic regression, decision trees, random forests, or gradient boosting machines. Split the data into training and testing sets, train the model, and tune hyperparameters to optimize performance. Evaluate the model using testing data and cross-validation. Deploy the model into production to make real-time predictions and prevent churn.

Use an automated predictive model

Do people with lower NPS scores tend to leave more? Are they evaluating competitor solutions? You can predict who might leave by spotting these signs in the data. You must use machine learning with a dataset containing all the information you have about customers who stayed and those who left. The algorithm learns from this historical data to understand how different factors relate to churn. Then, it can predict if future customers with similar behaviors might leave or stay.

💡Factors can help you identify customers evaluating competitor solutions by helping you track when they visit their G2 pages.

Establish retention strategy

Optimize your retention strategy by prioritizing actions based on the probability of customer churn. When customers first sign up, use checklists and personalized help to ensure they understand and use the product. As they keep using it, watch out for signs they might leave. For instance, if they're not using a feature they need, you can send them helpful tips to get them back on track.

Track results regularly

Continuously monitor the churn prediction model's performance and update it with new data periodically to ensure it remains effective as customer behavior evolves. Before introducing any further changes to your plan, collect enough data to measure the real impact of your efforts.

Your churn model will provide probabilities for various customer segments. It's essential to keep testing new strategies and recording the impact on these segments. While creating a mathematical model of churn requires time and resources from your team, each test can help you create a better model for the future.

6 Customer churn prediction best practices

Now that you know how to build a churn prediction model, here are a few handy tips you must remember to prevent customer churn:

1. Segment Customers

After obtaining your data, it's time to shift your focus towards the users and begin segmenting them. Since users have distinct needs and usage patterns, they don't churn for the same reason. Hence, it's essential to categorize them into separate segments. You can segment them based on their:

Demographics, such as location, region, company size, and the year they signed up for your company.
Behavior and usage, such as how often they log in, whether they use a particular feature more or less, or whether they have completed the onboarding process.
Contract terms include pricing plans and whether customers signed up for a monthly, quarterly, or yearly deal.

You can design retention strategies and marketing campaigns tailored to specific customer segments by segmenting customers based on their churn likelihood and characteristics. Domain knowledge or clustering techniques can help you define meaningful segments.

2. Monitor product usage data of existing customers

Product usage data captures how and when customers use your software. Important data points include feature usage, customer behavior, clicks, and other metrics such as time-to-value and product stickiness.

To build an effective model, it's important to consider some key product usage data points such as:

Monitor feature usage data to show users' engagement with your product's features, indicating popularity and relevance.
Observe users’ actions within your product, like when they use it, how long they use it, which features they engage with, and how they progress through it.
Track the number of times a user clicks or interacts with a UI element, such as a button, checkbox, text area, or menu.
Track other product usage data such as time-to-value, product stickiness, interactions, and more.

3. Keep an eye on customer success metrics

Understanding your users' success with your product is crucial in determining if they will continue using it. Surveys such as NPS and CSAT can be used to measure customer success. An NPS score of less than 7 or 8 means you may need a win-back campaign or value incentive to change their attitude towards your product. NPS measures loyalty and willingness to recommend, while CSAT measures customer satisfaction. Conduct these surveys periodically to track customer success and satisfaction.

4. Gather customer feedback regularly

Apart from gathering feedback through conventional ways, you can use various other forms of customer feedback to gain insights into their experience with your product or service. For example, in-app surveys can provide you with contextual input from users. You can use them to find out about your customer's overall satisfaction with your product, their experience with a particular feature, any issues they may have faced, or even the features they would like you to add or implement. This can be very helpful in understanding the general sentiment of users and identifying areas of improvement or strengths.

To promptly address issues and demonstrate responsiveness to user input, incorporate real-time feedback loops within your product. Acknowledge the feedback received through in-app surveys and communicate any actions taken to address user concerns.

5. Enhance customer experience

You can streamline the customer experience using automated onboarding, self-service options, and personalized support. Furthermore, you should use customer feedback to identify areas of improvement and proactively address any customer dissatisfaction rather than reacting after the fact.

6. Improve customer service

Respond promptly to inquiries and complaints, offer helpful advice, and measure performance using metrics like support tickets, call center response times, and social media interactions. Monitor these metrics to gain insights into customer service trends and effectiveness.

Customer Churn Prediction: Key Steps & Benefits

Predicting customer churn helps businesses retain clients and reduce acquisition costs through data-driven strategies.

1. Key Steps in Churn Prediction: Data collection, feature selection, model selection (Logistic Regression, Decision Trees, Random Forests, Gradient Boosting), model training, and real-time monitoring.

2. Essential Features: Customer tenure, usage frequency, service interactions, and engagement metrics.

3. Strategic Benefits: Identify at-risk customers, implement targeted retention efforts, and enhance profitability.

Leveraging churn prediction models enables businesses to proactively improve customer retention and long-term growth.

Wrapping up

Reducing customer churn is crucial for businesses as it directly impacts long-term revenue, customer loyalty, and overall business stability. However, understanding why customers leave requires analyzing data and tracking various metrics over time. Effective churn analysis involves monitoring overall customer turnover rates and identifying factors contributing to attrition.

Businesses can use advanced analytics and machine learning techniques to identify potential churners and implement targeted retention strategies.

Top 10 Lusha Alternatives for B2B Contact Enrichment

April 28, 2025

•

0 min read

Top 10 Lead Forensics Competitors for Visitor Identification in 2025

April 9, 2025

•

0 min read

Top UnifyGTM Alternatives for Signal-Based GTM

April 9, 2025

•

0 min read

Improve LinkedIn Ads Targeting with Audience Builder

March 20, 2025

•

0 min read

When you think of launching an ABM campaign, it’s natural to consider running LinkedIn ads. All you need to do is build and upload an audience list of your target accounts and launch your ads… right? Unfortunately, creating audience lists isn’t a walk in the park.

Your ability to maximize the value of your ads depends heavily on how you target them. If you want to build the perfect audience list, you need to look through mountains of data across multiple tools and Excel sheets while manually categorizing it according to your campaign goals.

Luckily, we have “Audience Builder” to prevent you from wasting your money.

In this article, we’ll show you how using Audience Builder is an effective way for you to target your audience on LinkedIn 🎯

Audience targeting: why the old way doesn’t work

Marketers typically build multiple audience lists for several different marketing initiatives, such as:

ICP lists
ABM lists
Retargeting lists
Webinar lists
Intent data lists
SDR lists, etc.

You might rely on LinkedIn’s native targeting filters to build your target account list. However, this comes with a set of challenges, such as:

LinkedIn native targeting filters don’t include intent data, making it difficult to specifically target high-intent accounts
LinkedIn’s job titles and industry tags are still being updated, so you might unintentionally target accounts that don’t fall under your ICP.

Of course, you can also build target account lists outside the platform via Apollo or Zoominfo and upload them to Campaign Manager. However, building out and vetting this list while sorting through different data sets would be time-consuming. Let’s say you want to build a list excluding your existing customers or those who have already signed up for a demo. It becomes tedious to scope out and filter every single account. Plus, you also have to continuously update this list every day according to when prospects book demos.

Now that’s A LOT ☠️

Abhishek at Descope explains why they faced a challenge when creating audience lists for their LinkedIn ad campaigns:

“Most tech companies have very complex buying groups with multiple stakeholders where the decision-maker and influencer differ. We were looking for a scalable way to craft the right message for the right buyer role at the right organization size without blowing the company’s budget.” – Abhishek Iyer, Director of Marketing at Descope.

For all these reasons, you need a tool that consolidates and integrates all your account data across multiple tools and automatically creates an audience list of accounts interested in your solution. Fortunately, your search ends here!

Introducing Audience Builder

Audience Builder allows you to build any audience from your Factors data and automatically sync these lists directly to LinkedIn. Let’s show you how:

Segment your audience lists based on intent data

Let’s say you want to show your ads specifically to accounts who visit your pricing page. You can use the event as a trigger to add them to a LinkedIn audience. Now, you can show ads to product-aware accounts, aiming to engage them further after they've shown interest.

You can skip the guesswork and leverage the power of intent data to target the accounts genuinely interested in your product. You avoid wasting ad spend and can have your GTM team prioritize the right prospects to meet your pipeline targets

Sync data from multiple sources

Every marketing team uses various tools across their tech stack, each with different capabilities. However, syncing data across these tools to create an airtight audience list can be tricky.

Let’s say you use Salesforce. Since Salesforce doesn’t have a native sync with LinkedIn, you have to manually sort and build the list before uploading it to Campaign Manager. Plus, Salesforce doesn't offer website intent data or G2 intent data, so you cannot filter accounts that are genuinely interested in your solution. That’s where Factors comes in.

We help you unify end-to-end B2B customer journey data from G2, Linkedin, CRMs, and MAPs, giving you insights into every customer segment, their intent signals, and how to reach them effectively.

Here’s an example of how Descope uses the Audience Builder feature to build their lists:

“Factors provides granular visibility into accounts visiting our website as well as their engagement across G2 and other channels. Accordingly, we receive tons of insight into account activity without the need for form submissions or sign-ups.

Based on this, we’ve created 3 audience segments: Top, middle, and bottom of the funnel. These audiences are pushed into LinkedIn campaigns and served tailored ads based on their sales cycle stage.” – Abhishek Iyer, Director of Marketing at Descope

Audience Builder use cases

Now, you can leverage every bit of GTM data to create and activate the most relevant audiences. Here are a few ways you can use this shiny new feature to supercharge your LinkedIn ads:

US-based software companies with 500+ employees that visited the pricing page >2 times and are not in the CRM: When ICP accounts visit a high-intent page like your pricing or demo booking page, they are far more engaged than just regular website visitors. You can create a list of such accounts in your CRM and launch a retargeting campaign where you can use ads to drive consideration for your product.

ICP accounts that clicked on a search ad but failed to convert: Suppose a set of accounts clicked on a search ad that includes a feature keyword. You can use this data to push these accounts to a list that shows ads related to the relevant feature. Plus, you can add a condition to exclude companies who booked a demo.

Enterprise Software and IT services companies that engage with more than 5 pages of your G2 category: When you notice accounts engage with your G2 page, they are evaluating your solution. Create a list targeting decision-makers and launch an ad campaign highlighting your best reviews and case studies.

ABM target accounts that viewed LinkedIn ads and then visited the website: Prospects that visit your site after clicking an ad are likely problem-aware, so you can create an account list and add them to a sequential ad campaign to move them further down the funnel.

Once Descope used Factors to build audience lists, they saw positive results and could easily target their ICP accounts without sifting through loads of data across their tech stack:

“Factors is already playing a key role in helping us make our LinkedIn paid spend more efficient. Even if one person from a specific account visits our website, Factors helps us target decision-makers and the larger buying committee as a whole to ensure that all the right people from a target account see our ads. Ultimately, this helps our LinkedIn ad budgets go that extra mile.” – Abhishek Iyer, Director of Marketing at Descope.

Enhancing LinkedIn Ad Targeting with Factors.ai's Audience Builder

Factors.ai’s Audience Builder refines LinkedIn ad targeting by leveraging intent data for precise audience segmentation.

- Targeting with Intent Data: Identify high-intent accounts, such as those visiting pricing pages.

- Automated Audience List Creation: Integrates data from multiple sources to streamline segmentation and reduce manual effort.

- Direct LinkedIn Sync: Ensures ads reach relevant prospects in real-time, boosting engagement and conversions.

By addressing limitations in LinkedIn’s native filters—such as outdated job titles and lack of intent data—Audience Builder enhances targeting accuracy, improves campaign performance, and drives higher ROI.

Join the waitlist today

Building a solid audience list is key to targeting your ideal prospects. With AdPilot, you can use data-driven insights to effectively target the right accounts at the right time and double your LinkedIn ROI. Speak to our experts today to learn how AdPilot can revamp your LinkedIn ads strategy.

LinkedIn Industry Tags 101: What Marketers Must Know

March 20, 2025

•

0 min read

LinkedIn is truly the place to B2B, isn’t it?

80% of B2B marketers say LinkedIn is part of their advertising strategy because 4 out of 5 of its 900 million members drive business decisions, making it a key platform for lead generation. Marketers can launch ad campaigns to target decision-makers from small businesses to Fortune 500 companies worldwide.

LinkedIn’s robust campaign manager platform allows companies to set their targeting criteria based on 20 different attribute categories, such as company, job experience, education, demographics, interests, and traits.

The image displays a portion of a LinkedIn Help page under the " Marketing Solutions " section.

However, while LinkedIn campaign manager is a boon for running B2B ads, there's room for refinement when it comes to the ad platform's industry tag categorization and audience targeting mechanism.

The LinkedIn industry list currently consists of 24 main categories and 148 subcategories as applicable industries for company profiles. These categories are presently visible for company pages but are yet to be updated on Campaign Manager.

The image displays a comprehensive list of industry sectors formatted in a table — **Source:** **The Linked In Man**

While these categories cover a wide range of industries, this article explores why they may still be insufficient — and how we can overcome the hurdle of vague industry tags to optimize ad performance ⬇️

How Does LinkedIn Campaign Manager Define Industries?

LinkedIn defines industry as the company's primary industry, which is where the member is employed, as stated by the company. Additional industries may be inferred about the company and included for targeting.

Individuals can’t choose the industry but rather get assigned the company's industry to which they are attached.

The problem arises when there is limited clarity on which industry a particular company belongs to. When selecting the industry option of a LinkedIn company page, the creator or page admin determines the industry. Since these are subjective, irregularities can occur especially when a company can come under two different industries.

For instance, a health tech company can come under “health, wellness and fitness,” “hospital and health care” or “software development”

Let’s look at this with a detailed example 🔽

Suppose you want to showcase your ad to decision-makers working in fintechs specifically. Here are examples of 3 fintech companies and how LinkedIn identifies their industries:

1. RazorpayXPayroll is placed under “IT services and consulting,” whereas it’s payroll software.

2. PayPal & Payoneer are similar platforms that facilitate international bank transfers but are under different industry tags.

As you can see, all 3 companies are virtually the same but are categorized differently on Linkedin. Seems confusing, right?

You risk losing out on ICP companies or worse you spend on irrelevant companies that are not your ICP because LinkedIn's categorization is different from your expectations

For instance, if you want to target fintechs and pick “financial services” in campaign manager, you’d also end up advertising to banks and investment companies.

Or if you pick “software and development,” your ads are shown to every other software company, regardless of whether they come under your ICP.

And we all know that an unqualified prospect can take a lot of time from your sales and marketing team, costing your company more than it pays.

Now the real question is,

How Do You Overcome This Problem?

Here’s what Tim Davidson, VP of Marketing at B2B Rizz, has to say:

As mentioned above, creating a target account list on a third-party platform allows you to present your ads to high-intent companies that actually fall within your ICP without overshooting your paid ad spend.

You can either build a list of cold accounts on a database tool like Apollo or ZoomInfo or build granular segments of warm ICP accounts engaging on your Website, LinkedIn, G2, CRM, etc. inside Factors.

💡You can use Factors account segments to identify and create a list of web visitors segmented by source and how far along they are in the customer journey. You can also refine the list by targeting accounts that visit high-intent pages (pricing pages, comparison blogs, G2 reviews, etc.) and fit your ICP based on demographics, industry, technographics, revenue, etc. Once done, you’ll have a list of high-fit, high-intent accounts.

Upload this list when creating audiences on LinkedIn to skip the ambiguity and save ad spend. It also comes in handy when launching retargeting campaigns to prospects in the solution-aware stage.

Understanding LinkedIn Industry Tags

LinkedIn industry tags help categorize companies into 24 main categories and 148 subcategories, enabling audience segmentation for B2B advertising.

- Challenges in Targeting: Categories can be ambiguous, making precise audience segmentation difficult.

- Classification Issues: A health tech company may fall under multiple categories like "Health, Wellness and Fitness," "Hospital and Health Care," or "Software Development."

- Optimizing Targeting: Combine industry tags with job titles, company size, and demographics for better audience reach.

Refining targeting strategies with additional attributes ensures more accurate segmentation, improves ad performance, and enhances B2B marketing effectiveness.

Wrapping Up

LinkedIn’s native targeting features while useful still have some room for optimization that the LinkedIn team is currently working on solving. In the meantime, you can use target account lists to save time and exclusively target your ads to prospects in market for your solution.

Find out how you can use Factors.ai for LinkedIn retargeting

And guess what? We’re coming up with something exciting that can help you revamp your LinkedIn ad strategy and make the most of LinkedIn. Stay tuned for more!