How Appcues Personalized their Sales Outreach with Product Triggers

A guest post by Ed Fry, Head of Growth @ Hull.io

If you’ve read the latest chapter in Data Driven Sales, you may have marveled at how DigitalOcean drastically improved their sales efficiency with trigger based sales outreach.

If you missed it, they improved enterprise conversion rates by 30%. Combining firmograpic data with product usage data, they were able to alert the sales team whenever a promising account had usage that indicated they might be ready to buy.

The logical next question – How do I leverage my own data to find better sales triggers?

And you may not (yet!) have the 500,000 customers to analyze across two data science teams...

Appcues is one of those companies.

John Sherer, Director of Sales was suffering a similar problem to DigitalOcean. With a trial-based sales model, they couldn’t sell a dream. They had to sell on the reality of what a trial user had already experienced.

The sales strategy is to identify good fit businesses and focus in on the trailers who were taking meaningful product actions. Appcues used Clearbit Enrichment data (company size and job title) to identify ‘good fit businesses’ and pass them to sales for additional qualification. But the missing piece was understanding when users were taking meaningful actions in the app. Specifically the actions that indicated that they were ready to buy or upgrade.

Sales people shouldn’t be in Mixpanel

The insights for this all lay in their analytics tools. But John didn’t want his sales team “bouncing between tools”, especially going through trial accounts one-by-one searching for a fit. “Sales people shouldn’t be in Mixpanel”, he told me. Instead, they should be in a focused workflow in the tools they're already using.

At first, John tried using a product usage score model. Instead of analyzing multiple analytics events, the idea was to aggregate this all into one score that sales people could use and could be easily synced between tools. This usage score compiled different actions like installing, setting up a workflow, and inviting users.

• Trial has expired
• Embed code is installed
• Activated an integration
• Invited a user
• Requested a CSV export
• Viewed subscription
• Published a set of hotspots in Appcues
• Published a modal

Appcues used Hull, a customer data platform, to take their analytics events and compute live scores for every trial customer. Those scores were then synced with their CRM.

In theory, this highlighted accounts which were a strong fit and more likely to buy. Sales could then see within HubSpot whether the account has a high usage score or not. No more bouncing between tools.

Except no one used it.

“With a small sales team, you can’t depend on a system to make the decision for the reps. The business is changing too quickly to lock into one process or one score. Instead, we notify the reps with information they can use to make intelligent and ever evolving decisions.

Instead of saying, ‘this lead is 66% more likely to buy’, reach out. We say, ‘this lead is a good fit and just installed’, engage. Over time, through iteration the reps can then best determine how engage for maximum results.”

With just a score to work with, all the context and reasons why a customer was getting value was obscured. There was no hint as to why an account was worth talking to.

Qualified leads are not enough to close deals.

For closing deals, it is not just WHO you talk to. It’s WHAT you say.

As DigitalOcean found, “let’s have a call” is not good enough. The same story was found at Appcues. Sales were cued in to the right accounts, but they didn’t have the “clues” to start a conversation. To them, the usage score directed them to be impersonal with qualified leads. Meanwhile, they had conversation starters with other deals they were working, particularly known accounts further down the pipeline.

Appcues needed to deliver both qualified leads AND a reason to reach out for sales to be truly personal and targeted. Lead scoring alone doesn’t deliver that.

Instead, Appcues packaged up their analytics data to notify their sales team when trial accounts were taking actions of interest. To do this, they used Hull to take the raw analytics events, and transform it into a format that worked with HubSpot and the sales team. This made sure sales always had a conversation starter.

But this still had something missing.

Sales wouldn’t be cued to take action in real time. Whilst CRMs can generate tasks and emails for sales people to take action, this wasn’t fast enough to take action on key accounts right away. Actions taken by the user are quickly forgotten by the user, which blunted the effect of this data-driven outreach.

Updated CRM profiles are not enough to close deals

For closing deals, it is not just WHO you talk to and WHAT you say. It’s WHEN you say it.

As well as enriching their CRM with all the “clues” that sales needed to talk with, they needed the timely notifications to prompt them to take action.

Appcues used Slack for internal discussions. The transient nature of Slack messages suited this task. If a notification was missed, nothing broke. It didn’t accumulate in an inbox or task list waiting to be completed. It simply becomes a feed of activity to take action on right now.

To package the data up to trigger Slack notifications, Appcues used Hull to create a segment of users with recent high value activities. Whenever a contact triggered one of those activities, they entered the segment and a Slack notification was triggered. The Slack notification would include a summary of their recent actions as well as their bio and key data from Clearbit. This gave sales the overview they needed to take action right away.

This enabled their small sales team to focus on the right accounts - and they knew WHAT to say, and WHEN to say it. Sales could speak directly to trial customers in a way that was both personal and timely - “Hey, I noticed you just installed. How can I help?”.

Moreover, each sales person had all these account lists and the reason for reaching out without leaving their workflow. As John says, “You’re giving a sales person a little more time back in their day” so they can do the job of finding the right people to speak to and saying the right thing without bouncing between different tools.

“There's an old sales maxim that says, 'You have 10 leads, 2 will buy even if you're not a good rep, 2 will never buy, even if you're fantastic, and your job is to win the 6 in the middle.' For now, we're fully focused on finding the 2 that will buy every time.”

Just as marketers use data to segment audiences and personalize content of emails, ads, live chat and more - sales needs to use data to cue in sales on the right accounts, and clue them in on what to say.

To do this, you need to package the data your sales team needs into a format their CRM and sales tools will inject, and then sync it across in a reliable way. Give your sales team a dependable feed of ways they can advance and close deals.

Moreover, once you’ve done the legwork to package your data up to send to your sales team, it becomes very easy to repeat for your other customer-facing teams like customer success and marketing - we call it Personalization Everywhere!