PPCLivePerson Says Bouncing is Single Worst Action a User Can Take

LivePerson Says Bouncing is Single Worst Action a User Can Take

LivePerson's latest product is Keyword Lift, which analyzes your paid search activity and suggests strategies to "intercept" users on your website to increase the chances of creating an assisted conversion via relevant content or instant chat

LivePerson introduce Keyword Lift

LivePerson may be most recognized for their instant chat software for websites which enables your customer service team to respond to users in real time whilst they are browsing your website, but this is just one piece of the suite of content marketing software that enables businesses to leverage customer engagement data in real time in order to place the right message in a timely manner via the right channels. LivePerson’s latest product is Keyword Lift, which analyzes your paid search activity and algorithmically suggests courses of action for the marketer to deploy strategies to “intercept” users on their website to increase the chances of creating an “assisted conversion” either via relevant personalized content or one to one human interaction via instant chat.

SEW spoke to Avinoam Zelenko, Product Owner of Keyword Lift and Analytics Driven Engagement at LivePerson, who explained that although this latest product is focused on paid search traffic, the software does not manipulate keyword bids (or change anything in your PPC interface) but instead is “an intelligence layer” of real time on-site performance metrics such as content clicks, page views per user, action types (e.g what a video, email sign up), bounce rates and exit rates to bring more color to the actual conversion process. Zelenko told SEW, “We apply predictive analytics to the website data to identify how to improve the conversion rates on ‘painful’ keywords which have delivering ROI under par.

Keyword Lift essentially suggests tactical responses and creates jobs for assisted conversion channels which the marketing manager can assign to customer service representatives. For instance, on keywords with low profit margins, the manager can activate a targeted custom content routine which is designed to echo the perceived intent. Whereas with high margin keywords, the manager can leverage text or voice chat to respond to the user on the site directly. As in the case of pay-per-call ads, depending on how complicated the buying decision or purchase process is, the transcripts and data from one to one conversations with a customer service representative can be a more profitable way to measure the effectiveness of paid search ads. Similarly, Keyword Lift should enable managers to reduce wastage around the associated costs of customer service representatives as jobs will only be assigned to users who have been identified as showing a high propensity or intent to convert.

LivePerson can echo the search query of the user with content, to make offers on similar products, or use instant chat so a user can be instantly request a product and be led to the correct product page.

LivePerson’s Keyword Lift system is a fully automated, ‘self learning’ machine which refreshes the predictive model every 24 hours. Whilst currently it is in “black box” mode Zelenko shared some insight into its inner workings, “The common misconception is that that there is a strong correlation between time on site and page views per users, when actually they is a lot of noise and it is not correlated to any kind of prediction of user behavior. At best, they are a kind assisted metrics.”

“What we actually found that bounce rate was the most highly correlated to predicting conversion. Bouncing is the single worst action a user can take on your website.”

Whilst this might seem almost too obvious to mention that the ‘worst action’ a user can take is when they take no action at all (whereas ‘exit rate’ is at least the user leaving for a reason – a ‘bounce’, which is calculated as a single page visit, means they did absolutely nothing), to recontextualize the problem in this way, makes the solution easier to identify. Zelenko continued, “We decided on a different approach. If bounce rate was the biggest predictor to user conversion, what if we can take an action to stop the user bouncing?”

Using custom content or engagement actions such as text chat, these site side actions to reduce the bounce rate also help to identify what ‘upper funnel’ engagements need to happen. Just as in-store signage around offers or deals or a shop assistant asking if you need help are engagement actions in bricks and mortar stores, LivePerson is trying to simulate the experience in a web environment.

Resources

The 2023 B2B Superpowers Index
whitepaper | Analytics

The 2023 B2B Superpowers Index

8m
Data Analytics in Marketing
whitepaper | Analytics

Data Analytics in Marketing

10m
The Third-Party Data Deprecation Playbook
whitepaper | Digital Marketing

The Third-Party Data Deprecation Playbook

1y
Utilizing Email To Stop Fraud-eCommerce Client Fraud Case Study
whitepaper | Digital Marketing

Utilizing Email To Stop Fraud-eCommerce Client Fraud Case Study

1y