The toughest sales challenge product-led companies face is figuring who to talk to in their self-serve funnel. At scale, it’s not practical to talk to everyone who signs up. On the flip side, if you just wait for customers to raise their hands, you’re missing out on opportunities. You’ll miss out on prospects who get frustrated trying your product or find a competitor they like better.
So to solve this, companies set up scoring rules to prioritize accounts. 10 new users in a week → +10 points. Created 25 projects in the first month → +25 points. Signed up from a freemail domain → -20 points. The problem with this approach is that it’s arbitrary and brittle. Your scoring is only as good as humans can reason about the data and Product-Led companies have too much data for humans to make sense of.
There’s a better way: machine learning.
A decade of ML experience
We saw the power of machine learning (ML) first when we worked at Sift. Fred, my co-founder at Calixa, was also the co-founder of Sift. We used ML to catch fraud on the internet for companies like Twilio, Box, and McDonalds.
We built systems that detect fraud in real-time across over a billion users per month. Our ML would look at thousands of different signals to figure out who could be trusted. Fraud is really tricky because you have criminals on the other side trying to evade whatever systems you put in place. So our ML needed to continuously evolve as fraudsters devised new attacks.
We often had companies say to us, “you’re great at catching our bad accounts, how about helping us find the good ones.” And it made sense. We built models to detect fraud based on user behavior. Why not look at user behavior and find the good accounts.
When Fred and I brainstormed the next problem we wanted to solve, it was easy. We saw the massive untapped potential that Product-Led companies have in their self-serve funnel. They had all this rich data but couldn’t do anything with it. We knew that we could use machine learning to help companies turn sign-ups into revenue.
Enter AI-Powered Prospecting
Product-led companies are sitting on a goldmine of data.
Traditionally, companies would just have access to firmographic data. Which company someone was from. What their title was. What their lead source was. You’d use this data to prioritize your outreach.
In the era of product-led, accounts are generating a rich set of data beyond this. Accounts are using your product and showing clear signs of intent. This intent data is the unique advantage that product led companies have over a traditional GTM motion. But how do you tap it? If you don’t have a dedicated data science team, you’re stuck.
So that’s why we built AI-Power Prospecting. We are empowering Sales to take control of their data. Calixa finds the accounts your reps should be talking to and gives them insights on what to talk to them about. Our machine learning models leverage both your product usage and firmographic data to find the most promising accounts.
Our ML is not a black box. We give reps both a score and a set of reasons to understand what signals our systems have picked up. Those signals can then be used to run playbooks or drive conversations with customers.
The best part is that it’s more than just a score. You can use it in any of your Calixa workflows and automations. Alert reps via Slack to a new hot lead. Automatically add a lead to a sales cadence or marketing campaign. Surface the data inside of Salesforce to give reps extra context. All of this is possible and more.
Machine learning made easy
To teach our AI, it's easy. You just tell us which of your existing accounts are examples of ones you want us to find. Our systems then take those examples and build models unique to your business and customers. We figure out which behaviors and attributes are indicative of your best accounts and then look for those signals in your other accounts.
What’s even better is that we can build models for multiple segments of your business. We first start with finding your product qualified accounts but can easily extend this to predict other things like upsell opportunities. Calixa helps sellers across multiple phases of the customer's buying journey.
The machine learning that powers our AI-prospecting is trained daily so that it always stays fresh. Because we keep training your models, we can adapt to changes in user behavior. For example, if you were to add new features to your product or add a new feature gate, our ML will adapt. It will learn the new patterns of good customers.
Most importantly, we continuously test our predictions to see how we did. This helps us catch cases where we were wrong and can feed those learnings back into the model. We can even take signals from how your reps are engaging with accounts as feedback to improve our predictions.
Powering Product-Led Sales
We built our AI-Powered Prospecting to give Sales control of their own destiny. We kept hearing from Sales teams at Product-Led companies that they were not getting access to the data resources they needed to run their GTM motions. With Calixa, it’s now easy to get started running Product-Led Sales motions at scale. We’ve done all the heavy-lifting so you don’t have to.
If you have feedback for me, just shoot me a note at firstname.lastname@example.org. I’d love to chat.