Does Chat Result in More Good NET LEADS?

A Group – 50% of Traffic
Phone Tracking = On
Form Tracking = On
Chat = On

B Group – 50% of Traffic
Phone Tracking = On
Form Tracking = On
Chat = Off

RESEARCH:  When holding other variables constant, what is the impact on net qualified lead generation of on-site, human-powered chat compared to no chat, in professional-services businesses (law firms, plastic surgeons, roofers, retirement homes etc) servicing the B2C market. Especially when website visitors have other options, including calling an on-site phone number or filling in a lead capture form, to reach out to the firm in question.

Executive summary:  Many websites exist primarily to capture leads.  We seek to understand how to maximize conversion rates through use of different conversion methods, primarily through introducing human powered live chat to see how it affects both phone leads, form leads, as well as overall qualified lead generation.  This research was seeking to discover if on-site live chat would significantly outperform no live chat at all in the net number of qualified** leads generated.

**A qualified lead is an inbound request that the firm manually confirms there is a chance of the firm being hired. Increasingly marketing is driving higher volumes of inquiries/”leads”, which look great from a metrics/tracking perspective, however until that inquiry is qualified it is a vanity metric. Therefore an additional layer was added to filter out qualified, from unqualified, leads.

Purpose: The purpose of this research was to take a range of websites representing different B2C professional services businesses and see how on-site live chat impacted the generation of qualified leads. 

Other criteria included:

  • A mixture of traffic generation methods including organic search, social, paid social, paid search and others
  • A reliable traffic source producing around 50 hits per day per website
  • A mixture of geographical markets
  • A mixture of client services

Methodology: Once a website was selected, and the team agreed to be part, every website was then provided an A/B testing scenario with half the traffic to the website being offered proactive on-site live chat option (manned by Juvo Leads agents) and half the traffic would not have live chat as an option. 

All websites would maintain a prominent phone number call to action and lead capture form. This way each site would act as a control group for itself and the research was effectively a collection of separate A/B tests.

Once the test began the process of collating raw, unqualified numbers was automated. Anytime a call was made to the listed number, anytime a form was filled in or anytime a visitor “chatted” with the JuvoLeads agent – these were aggregated to a total of 311 not-qualified “interactions” across the cohort.

This broke down to aggregate numbers of:

  • 115 First Time Callers
  • 74 Lead Forms
  • 122 Chats

With total interactions (not-qualified leads) for Group A (with chat) of 201, broken down as follows:

  • 49 Calls
  • 30 Forms
  • 122 Chats

And total interactions (not-qualified leads) for Group B (no chat) of 110, broken down as follows:

  • 66 Calls
  • 44 Forms
  • 0 Chats

The raw difference is instantly noticeable with Group A (with chat) generating 82% more interactions than Group B (no chat).

Because these numbers are raw, they don’t show the true picture. To delve deeper each interaction was scored as to determine the true number of qualified leads (potential customers  that may end up hiring the firm). 

To gain qualification data for phone and form leads each of the firms had to manually provide this data. In all:

  • Of the 115 calls, 28 calls (24%) were deemed qualified
  • Of the 74 forms, 22 forms (29%) were deemed qualified
  • Of the 122 chats, 31 chats (25%) were deemed qualified

The most interesting thing however was to aggregate the three methods together to see how chat impacted the number of qualified leads:

This data is collected from 18 websites. The research period was 8 days. The total number of website visitors was ~7,000 visitors with a traffic split of 51% in Group A (with chat) and 49% in Group B (no chat).

Looking purely at the top line increases in total qualified leads, 53% more leads in Group A (with chat) than Group B (no chat) it’s easy to see that across a large enough sample size websites with on-site, human-powered chat significantly outperform websites without chat in the B2C professional services market.

It can also be noted that chat is one of the most effective methods of initiating interactions with visitors to a website, as is evidenced by the 82% increase in interactions with Group A (with chat).

Then combine that with the factor chat is more time efficient than calls or forms; any given team member can handle a single call at any given time, whereas a chat agent can handle more than one interaction at any given time and forms are a lagging method of communication that naturally result in lost interest when the firm does get around to making contact whereas a chat is interacting with potential leads while they are already on the website.

Therefore it can safely be said: When holding other variables constant, on-site, human-powered chat outperforms no chat, in professional-services businesses servicing the B2C market.

Appendix A: 

Pipeline Revenue from Qualified Leads

As part of the follow up process additional information was gathered and one of the more interesting pieces of information was the difference in total “pipeline revenue” was generated from the two scenarios.

Pipeline is defined by each of the participating websites individually, and manually, assigning a dollar amount to each of the 82 total qualified leads as each lead is spoken to. Then splitting these leads into the various “With Chat” and “No Chat” cohorts. 

As you can see in the below chart, the “With Chat” cohort significantly outperformed the “No Chat” in absolute terms. When factoring in that the traffic split was 51%:49% we can determine that “With Chat” is a significantly more efficient use of traffic to generate Pipeline Revenue.

Further research may be conducted to breakdown these numbers in greater detail at a later date.


As the main focus of this research was purely on the qualified lead generation and the process of taking a lead to revenue requires a collection of additional and varying systems, processes and procedures, it was not within scope to determine how much of the above pipeline was realized.

Done For You A/B Test Trial

Sign up for a demo today let us know you want to a “Split Test Trial”.  We give qualified agencies “Split Test Trials” that shows the impact of chat on total leads captured.  Sign up today to get started!