How Conversational Marketing Skyrocketed My Website’s Conversions by 370%

In this blog post, Lukas Mignoli, founder of an online e-commerce store specializing in customizable t-shirts, details how, as a solo entrepreneur, he managed to skyrocket conversions on his website by implementing an effective AI-powered conversational marketing strategy, leveraging Callin.io. Source: medium.com.

I have experimented with various marketing strategies to grow my online business sales, including SEO and advertising.

An underutilized channel by solo entrepreneurs is customer care.

As Gabriel Weinberg and Justin Mares highlight in their book “Traction, a Startup Guide to Getting Customers”, customer care and sales are among the best marketing channels for startup, a sentiment I share.

Implementing a live 24/7 sales and customer care team, especially for a global business like mine where potential traffic comes at all hours, can be costly.

That was until I found an automated conversational marketing solution not just for e-commerce owners but for anyone with an online business selling goods and services through a website.

I installed and tested a WordPress plugin that transformed how I interact with site visitors, leading to a conversion increase of 370% in just four weeks.

This conversational marketing tool allowed my custom t-shirt e-commerce to leap from 2,800 USD to about 13,000 USD in monthly revenue.

Here’s how I did it, through technical adjustments and continuous optimizations…

Introducing the Conversational Marketing tool

Callin.io is not just any marketing plugin.

It promises to enhance interaction with website visitors by offering real-time callbacks (within 15 seconds) for phone consultations or more information on products or services.

However, the one making the calls – and this is the main peculiarity – is not a telephone agent or a flesh-and-blood consultant, but rather a conversational marketing bot, based on artificial intelligence and specifically trained to have a telephone conversation with website visitors, to convert them into potential customers, using conversational marketing techinques.

My goal was to provide a constantly active, efficient, and high-quality telephone consultation service, giving users detailed information and answers to their queries before making a purchase.

Despite initial doubts about handling a large volume of calls, the plugin, as stated by its developers, can manage more than 1 million calls simultaneously—an impressive feat for solo entrepreneurs like me, allowing for a 24/7 sales consultant availability without a dedicated team.

Introducing the Conversational Marketing tool

Setting up and configuring Callin.io was straightforward.

After downloading it from the WordPress plugin search engine, I installed this powerful conversational marketing on my website and configured my first AI phone bot by simply entering the API key.

Instructing the Algorithm

To instruct the AI on call conduct, topics, and user approach, the tool allows customization of prompts and greetings.

Additionally, the Retrieval Augmented Generation (RAG) function is indispensable for enabling the AI algorithm to learn how to speak during phone conversations.

To do this I used as ‘source of knowledge’ the entire text description of my website catalog and previous call transcriptions that I had with real leads and customers.

Through RAG, I optimized the bot’s performance by uploading entire phone conversations, significantly improving its efficacy.

Graphic customization

The next step was customization.

The plugin allowed me to adjust the pop-up’s message and appearance to match my site’s design, creating a cohesive user experience.

With the API key inserted, the system was ready to provide real-time AI-driven callbacks.

Initial results of my first Conversational Marketing experiment

The impact on my conversions wasn’t immediate. After activating the plugin, the first wave of results showed a 30% increase in conversions – a promising sign that prompted me to further explore optimization potentials.

The Breakthrough through A/B Testing

Aware that there was substantial room for improvement, I decided to implement a series of A/B tests, focusing on the greeting formulas and the prompts provided by the AI.

The goal was clear: to identify the combination of elements that maximized user engagement and, consequently, conversions.

For a period of 30 days, I tested various approaches.

In the first test, greetings ranged from an informal “Hi! Do you need help with our products?” to a more direct “Good morning, which products are you interested in purchasing?”.

Similarly, I configured various prompts with different tones and content, aiming to balance friendliness and the precise definition of the role assigned to the AI.

The prompt I tested in the first week was: “You are a sales agent. Your role is to assist customers in finding the perfect products to meet their needs, providing expert advice, and ensuring a seamless purchasing process from browsing to checkout“.

After conducting an A/B test for an additional 5 days, I decided to use the following prompt: “You are a telephone consultant for purchasing t-shirts. You work for the e-commerce. Your role is to guide customers in buying the product best suited to their preferences for fabric, color, model, etc.”.

This prompt allowed me to achieve a peak in conversions.

The most effective greetings I discovered were:

“Hello, I am your e-commerce consultant and I’m here to improve your purchase experience by providing personalized recommendations, answering any questions you may have, and guiding you through the entire shopping journey from start to finish“.

I must preface that these results are the fruit of continuous optimization work and must obviously be adapted based on context, business sector and on your conversational marketing strategy.

Although everyone has a natural intuition, no one can truly predict exactly what will work and what won’t! Therefore, I advise you to conduct A/B tests as I did.

Cost and revenue analysis

This experiment, in addition to being an investigation into conversion dynamics, also translated into a detailed cost-benefit analysis.

In just four weeks, I recorded an increase of 283 conversions. For the 283 sales generated, the total seconds of phone traffic used was 65,939, (equivalent to 1,098 minutes).

Considering the cost of phone traffic charged by Callin at 0.13 USD per minute, I invested a total budget of only 142.86 USD (there is no cost for installing and using the plugin and for creating a voice bot: you only pay for the phone traffic at the rate of 0.13 USD per minute).

However, the earnings spoke for themselves: compared to the previous month, I recorded an increase of 283 sales, with an average order value of about 36 USD.

This resulted in an additional gain of 10,296 USD per month, compared to the 2,849 USD generated in the previous month from 77 sales, with a slightly higher average order value of 37 USD. The total gain was 13,145 USD per month.

This conversational marketing tool worked really well!

Data Cross-Reference

It’s worth noting that I conducted a specific ex-post analysis to calculate how many people who had a telephone conversation with the phone bot actually purchased a t-shirt from my e-commerce.

Doing this analysis was much simpler than I thought: I just had to cross-reference the data.

I indeed connected the telephone numbers of leads who had a conversation with the AI with the “phone number” field entered by users when filling out the online order form in WooCommerce.

And the results turned out to be quite interesting…

About 27% of users who received advice from artificial intelligence proceeded with placing an order on the website.

Interaction with website customers

I then made another interesting discovery…

I personally called all the website users, not just the customers, but all those who had interacted with the bot to find that in 55% of cases, they did not realize they were communicating and interacting with an AI!.

Communication with the developers

Despite the extraordinary results, Callin.io still lacked a dedicated function for A/B testing (I conducted the tests by manually changing the rotation of the various bots).

This gap prompted me to directly contact the developers to suggest its implementation.

In my email, I outlined how the introduction of such a feature could not only simplify processes like the one I undertook but also further elevate the plugin’s performance, transforming it into an even more powerful tool for communication and conversational marketing.

I encouraged other entrepreneurs to do the same, convinced that a collective request could accelerate the introduction of this functionality.

Observations and future prospects

Switching from managing customer requests via email and WhatsApp to using Callin.io was transformative.

This experiment underscored the value of technological innovation and customer experience investment, proving that significant impacts can be made without large budgets.

With AI integration, the potential for further enhancing online business efficiency and customer service is immense.

Advice for Digital Entrepreneurs

Whether running a solo business or part of a larger organization, integrating this kind of technology can redefine customer interaction and boost conversions. Implementing this plugin was a game-changer, turning my website from a static display into an interactive conversion engine.

This conversational marketing tool represents a paradigm shift in e-commerce, combining AI sophistication with personalized engagement to meet today’s consumer needs.

I encourage everyone to explore the untapped opportunities in AI-assisted conversational marketing sector.

Notes from the CEO of Callin.io: We are truly delighted to learn how Lukas Mignoli, founder of an online e-commerce store specializing in customizable t-shirts, has significantly increased his revenue and customer engagement through our effective conversational marketing tool, Callin.io. We extend our best wishes to Lukas for his continued success and are pleased to confirm that we have gladly accepted his proposal to integrate an A/B testing feature, which we are currently developing. Thank you.