Executive Summary
- Fluke Health Solutions saw a 10.3x return on their investment (ROI) with Lift AI in the first year.
- Before Lift AI, the Intercom implementation of Fluke Health Solutions provided a new avenue for engaging with customers, but the sales team were having too many conversations about service and support issues, and wanted more quality leads.
- Fluke Health Solutions were relying on out-of-the-box features such as general site-wide greetings and page-based rules to engage visitors, resulting in low coverage and missed opportunities
- Lift AI revealed that only 6.3% of high intent visitors were going to the pricing page, which Fluke Health Solutions had set up as their primary chat strategy before implementing Lift AI.
- Using Lift AI, Fluke Health Solutions flipped their targeting strategy to be ‘buyer intent’ based. Lift AI uses a machine learning algorithm to score each website visitor’s likelihood to buy or convert, which is reflected by a ‘buyer intent’ score.
- With Lift AI, Fluke Health Solutions could take advantage of high intent visitors across their entire site.
Together, Lift AI and Fluke Health Solutions developed custom playbooks based on the unique buyer intent score of each visitor, which increased the average revenue per Intercom conversation by 246%, and increased revenue per site visitor by 345%!
Fluke Health Solutions consists of Fluke Biomedical which is the premier, global organization providing test and measurement equipment and services for the healthcare industry, and RaySafe, the leader in X-Ray radiation dosimetry management for radiographic imaging rooms.
Like many other forward-thinking companies, Fluke Health Solutions implemented the Intercom online chat and messaging solution in August 2020 to help generate more conversions.
Fluke Health Solutions knew that the biggest website opportunity was existing traffic, and Intercom enabled a toolset to target, engage, and convert those visitors in real-time, before they left the site.
It’s a smart move. It’s hard enough as it is to bring traffic to a website, so to lose out on those visitors after they navigate the site is a big missed opportunity.
Implementing a chat program has helped Fluke Health Solutions bring in new incremental revenue. However, they learned the same lesson that every other company learns in the process:
Impactful conversational marketing is hard to do. Really hard.
From website coverage to staffing, targeting strategies, page-based rulesets, chatbot and playbook design, and more - there’s a lot to get right. Not to mention that even with all of those aspects looked after, it can still result in frustration.
Why? Because Fluke Health Solutions wasn’t sure exactly which visitors to engage with Intercom. Essentially, it was a guessing game. Sales agents would engage with visitors just looking for support with existing products, or chatbots would engage visitors who were close to buying - but they got stuck in “bot jail” instead (when the visitor is chatting with a bot but really wants a human).
After bringing in Lift AI to help in February 2021, Fluke Health Solutions was able to increase the average revenue per chat conversation by 246%. Here’s how:
Figure out buyer intent first, then deploy playbooks second.
The problem with most chat implementations is that companies focus too much on complex “page-based” playbooks. These rules and criteria might include “engage the visitor with a live agent if the visitor is on the pricing page” or “engage the visitor with a chatbot if they’ve been on the website for more than 120 seconds” and a number of other rules.
The problem with this approach is that marketers and website owners are guessing what the visitors need based on hypotheses, which don’t accurately reflect what the visitor wants.
For example, if the visitor is on the pricing page, they might already be about to purchase and a chat invitation may actually get in the way of that (sabotaging the sale). The visitor could also be a competitor looking at a pricing page daily to see if there are any changes.
The other problem is that it reduces website coverage significantly. If a website is only engaging chat with visitors that meet the rules criteria, it might only be instigating chat with a tiny percentage of all visitors that have high intent. That leaves a lot of opportunity on the table.
But, companies can’t bring live chat to every visitor, right? If they did, sales agents would end up wasting a lot of their precious time chatting with the wrong visitors instead of making sales.
Companies also don’t want to engage every visitor with a chatbot, as they might frustrate the visitor with an experience that reduces the chances of getting a sale, or gets those visitors stuck in bot jail.
So, what’s the answer?
The answer is buyer intent - the likelihood of a visitor to purchase (or convert) on a website.
If companies know how likely each individual visitor is to convert in real-time, then they can engage visitors with a playbook designed specifically for their intent level.
That said, figuring out buyer intent in real-time is no simple feat. Only a machine-learning (artificial intelligence) model can do it, such as Lift AI.
Lift AI’s model was trained on billions of data points to predict the buyer intent of website visitors with incredible accuracy. By using historical data points plus real-time behavioural analytics, Lift AI is able to assign a buyer intent score to every website visitor on the Fluke Health Solutions website.
The beauty of this approach is twofold.
- It works in real-time, meaning companies can engage visitors with an optimized playbook based on their buyer intent scores before they leave the site
- It works on every single visitor - including those who are completely anonymous to the company’s ABM or CRM tools
Examples of winning Lift AI + Intercom playbooks
Lift AI assigns each visitor a buyer intent score between 0 and 100, where 100 represents a high intent to convert and 0 is a low intent. As visitors are scored, they are segmented into cohorts.
Lift AI then helped Fluke Health Solutions develop playbooks for each cohort which involved custom messaging and routing to either live agents or chatbots. For example:
- High intent playbook (live agents are online - send visitors to live agent)
- Medium intent playbook (live agents are online - use a mix of escalation chatbots and live agents)
- Low intent playbook (live agents are online - use chatbots only with narrow openings for escalation to live agent)
- Offline playbooks
- International geography playbooks
- + Many more
Here are a few examples of real playbooks that were set up:
Figure 1.0 - High Intent Targeting Example
Below shows the targeting set up inside of Intercom for a “Lift AI High Intent” visitor:
- Includes Lift AI scores between 60-100
- Excludes PPC campaign pages (instead, custom chatbots were developed for those campaigns)
- Fires Intercom within office hours (for live agents)
- Fires to visitors inside of North America
Figure 2.0 - Mid Intent Playbook Example
Once the targeting criteria are met, a playbook will be executed. Below is an example of a “Lift AI Medium Intent” playbook, which opens the conversation with a visitor (also known as a “send”). The opening message is important to customize based on the known intent of the visitor.
Figure 3.0 - Engagement Metrics for Fluke Health Solutions + Lift AI
Below is a breakdown of the number of sends, engagements, conversations, and leads created by each playbook. By tracking each playbook accordingly, Fluke Health Solutions and Lift AI were able to optimize their staffing and conversion rates accordingly.
Figure 4.0 - Old way vs New Way of Targeting
Unlocking the 94% of missed high intent visitors for Fluke Health Solutions
Once Lift AI started to gather buyer intent scores for each visitor, it became obvious that the previous strategy of focusing on the Fluke Health Solutions pricing page was leaving a lot of potential revenue on the table.
In fact, only 6% of high intent visitors went to the pricing page, meaning 94% of high intent visitors were on other pages of the website and were being missed.
This describes the problem of low “coverage” and the inaccuracy of the human hypothesis that visitors on the pricing page must be close to converting.
In reality, customer journeys are far more complex with many other variables that could predict their intent.
Adapting Strategies as Informed by Data
One of the other nuances of chat is the limitation of each toolset. For example, Intercom has restrictions on how many invitations can fire to each visitor in their journey. So, if Lift AI’s “low intent” playbooks were firing to relevant visitors, but those visitors then progressed into medium or high intent visitors, those visitors wouldn’t trigger the more valuable medium and high intent playbooks.
So, Lift AI adjusted the firing rate of low intent playbooks to ensure it was capturing as many medium to high intent visitors as possible within the limitations of Intercom.
The result? 345% Increase in Revenue Per Visitor
With the combination of intent-based playbook design, increased coverage, and customized messaging, Fluke Health Solutions was able to increase their average revenue per chat conversation by 246% and revenue per website visitor by 345% using Lift AI.
These results are typical of the Lift AI model. For example, customers using Drift get an average of 9x more conversions from chat into pipeline.
The breakthrough in AI technology allows marketers and sales teams to take advantage of the biggest opportunity on their website - visitors who are on their websites right now.
If you want to see results like Fluke Health Solutions, you can. Simply sign up for a free trial of Lift AI. The code snippet will be sent to you via email, and after a simple installation it will begin scoring your visitors. After 30 days, the team at Lift AI will give you a free revenue assessment to forecast your results and ROI.
