It’s no real surprise our team gets plenty of questions about AI email marketing – mainly because Mercanto invented drag-and-drop AI email marketing.

Getting started

On one side, we build a Spotify-style taste profile for each consumer.  On the other side, we use machine learning to model the attributes of each product in your catalog.  Then we take these two things and factor in newness, popularity, and diversity to recommend products that are fresh and relevant for each consumer.

Once you’ve installed Mercanto’s asynchronous tracking code on your site or in your app, our A.I. tracks and analyzes your visitors’ behavior. Meanwhile, we put a feed of your products through our natural language processing and semantic analysis algorithms. Lastly, our machine learning algorithms process all of this data in real-time to automatically populate the exact email content each user wants to buy, increasing sales and inspiring return visits.

Integration typically takes approximately four to eight engineering hours, depending on the complexity of your site. It’s a straightforward process, and we will have a customer success engineer walk you through every step and answer any questions you may have.

Yes, Mercanto’s technology is entirely ESP agnostic, and therefore we can work with any ESP. All you need to do is customize the variables that the ESP uses for the user/subscriber ID and the campaign ID.

Mercanto’s output is simply a snippet of HTML. You can paste that HTML into the email, and send it out as you usually would through your ESP. No custom integration is required.

For the purposes of email personalization, there are two broad approaches: collaborative filtering and content-based filtering.  Each approach has its own strengths and weaknesses, but both allow the marketer to automatically rank products.  This results in better personalization, faster campaign deployment, and higher ROI.

For email newsletters (- the workhorse of email marketing), content-based filtering is typically the better option because it creates a nuanced picture of every user’s tastes and interests, making it possible to show each one a wide variety of content that’s both relevant and compelling.

If you can understand the consumer’s interests and the product catalog at a deep, semantic level, you can deliver more relevant content and thereby create a more serendipitous consumer experience. 

We also use a combination of other best-of-breed AI technologies including neural networks, multi-word-entities, and semantic distancing. 

For example the ML knows that raincoat is more similar to umbrella than shoes. And also that raincoat is more similar to rain shoes than running shoes. 

You can learn more in this ‘Welcome to the AI revolution‘ playbook, or feel free to get in touch. 

The Mercanto onboarding process consists of four steps:

Step 1 – Kickoff meeting Your Account Manager will contact you to agree a mutually convenient time for the kickoff meeting.

Step 2 – Data integration and algorithm configuration When the data integration is done, Mercanto will adapt the algorithms to the specifics of your business model. 

Step 3 – Create content bricks for use in the drag-and-drop editor, and final QA After the data has been validated, the final step will be to create the content bricks (- the AI-driven content bricks) that are shown in the campaign. Your Account Manager will work with you throughout any final setup steps and testing to confirm everything is working properly.

Step 4 – Go Live! With all the preparation, on-boarding, and testing completed, you’re ready to go live and start using Mercanto every day! But that doesn’t mean our job is done:

  • Your Account Manager will be on hand to make sure Mercanto is continuing to meet your needs
  • Technical Support is available to help answer questions and solve any problems you may run into.
  • At any time, you can contact us at support@mercanto.app, via the help function in the platform, or any other way.

Implementation timescales This entire process will typically take approximately 4 weeks, depending on the complexity of your data.

This depends on two factors:

  • How often people click and open emails
  • Website engagement traffic

The more engagement, the faster you can begin making recommendations. As a ballbark figure, we’d suggest a month ‘listening’ to consumers before making recommendations.

The Mercanto platform can also use instore and in-app data, if that would be helpful.   

We work across all verticals here at Mercanto. We primarily focus on the retail, e-commerce, travel and hospitality sectors.

Yes. We work with all shapes and sizes of teams, whether it be the end client, their agency, the ESP and even other third parties. The Mercanto platform is actually built for the client so that any marketer can log in and easily create snippets of HTML within the drag-and-drop interface.

We also have support here within the business known as our Client Services team. The Client Service Team, together with our Data Science team, help you get up and running with the platform and act as an extension of your existing team.

Not a chance. Our data scientists tailor algorithms for each client. We consider your unique KPIs, business strategy, data structures, and content types, then create a solution that aligns with your model. Our algorithms continue to optimize and improve as your content and your customers’ preferences evolve.

Absolutely! Unlike some of our competitors, we do not rely on metadata or error-prone manual tagging. We believe your content is best understood through semantic analysis and insights, not manual tags. Once Mercanto is up and running and you’re sending us a feed of your content, our process automates the rest.

It assists immensely with quickly scaling your personalisation efforts. Individualised recommendations from machine learning have been shown to add 30-80 percent in email revenues for leading brands.

Key features

If a contact arrives on your site from a tracked link, a cookie will be set identifying the referring message ID and contact ID.

If the user has not arrived from a tracked link, the user can be identified from another method (Ex: site login, or sign up). You may set the identity of the user as a contact in the cookie.

If there is no referring message or known contact identifier available, an anonymous ID will be assigned to the session user. All event activity will be tracked and attributed to the anonymous ID. In the event that a contact identifier becomes available for the anonymous user, all subsequent activity will be tracked and attributed to the customer, and previously generated event activity from their anonymous session will be copied into their customer profile.

The product feed should include your product inventory, such as product image URLs, product names, SKUs, descriptions, price, and so on.

Depending on your business model and the size of the catalog, this feed may need to only include changes on a nightly basis, rather than a full sync each day. The feed can also be synched more frequently, for example on an hourly basis.

The Mercanto system is designed to support retail/ecommerce brands with anything between 30 and 1.5 million SKUs.  It works best when there are lots of consumers and lots of SKUs.

Our demo system holds 1.1 million SKUs. 

When inviting a colleague to collaborate in the Mercanto platform, you’ll want to make sure you give them the appropriate access:

Standard Users – get access to all tabs and features except the ability to add/edit team members under Settings > Team and make changes under Settings > Advanced. This is the most common role for anyone who is not the administrator.
Super User – is the supreme being. Has access to everything. This should probably be just one person, though an additional Super User as failsafe also works.
To create a user:

Navigate to Settings > Team > Invite
Add the user’s email and select their role.

Editing team member details or removing them: If you want to edit the email ID of a team member or if you want to delete a team member altogether, navigate to the ‘Team’ page after clicking on Settings from the top right hand button on the dashboard.

Yes, Mercanto has the proven ability to real-time rank catalogs with millions of items within milliseconds.

All products, long-tail low traffic, and new items will be displayed as Mercanto explores & ranks content based on the customer’s interests. The Mercanto natural language processing engine extracts attributes at the product level (within your catalog) and uses this information to make recommendations. 

The Mercanto API makes it easy to embed real-time ranked product for each unique customer on any page of your website. And it’s responsive for the mobile web.

At Mercanto, customer trust is our top priority.  Mercanto  implements responsible and sophisticated technical and physical controls designed to prevent unauthorized access to or disclosure of customer content.

We also rely on AWS to help boost our own security: AWS provide a data center and network architecture built to meet the requirements of the most security-sensitive organizations.

Can’t find what you need? Our customer care team is here to help.