Algorithm - Product Curation

Highlighted
New Member
10 0 0

Is there an app that performs like an algorithm? It can be composed of 2 or more data sets; 1) product data, 2) customer profile. When a client requests a product, the selection is narrowed down and ranked using the data sets. The optimum product is curated for the client. If there is no app, can a solution be engineered by a Shopify Expert?

0 Likes
Highlighted
Tourist
13 0 0

Hey Doug,

 

To clarify, are you looking for a solution within your site that can recommend products based on some selection/activity by a customer?

 

We have a way to do product recommendations for customers, but it's not a live on-site widget yet. Let me know if you'd like to discuss further! t@tresl.co

 

Tony

0 Likes
Highlighted
New Member
10 0 0

Not quite. Here's an example; Data set 1 contains product info on  100 shirts, e.g. - size, style, color, material, brand, etc. Data set 2 contains client profile info, e.g. - sleeve length, collar measurement, chest measurement. The client submits a request for a blue stripped dress shirt with a button-down collar. The algorithm searches data set 1 & 2 and recommends the single best match from the available inventory (in data set 1).

0 Likes
Highlighted
Tourist
13 0 0

Ah yea makes sense. So broadly there's two approaches to this:

 

1) Business rules. We create a comprehensive set of rules that encompasses all of the different styles and "paths" that a customer could go down to get a recommendation. So if a customer was looking for blue shirts, we would match all of the customer measurements + blue + shirt type, and return whatever comes up.

 

2) Machine learned. If we have enough data, we'd train a model on whatever customers with certain attributes seem to buy, and use those to recommend products. We can of course incorporate business rules here as well to constrain the recommendations so they aren't too farfetched.

0 Likes
Highlighted
New Member
10 0 0

That's intriguing. Have you created a similar solution within Shopify's platform? If yes, is there a sample of your work I may access? Thank you for contributing to my thread, it's greatly appreciated.

0 Likes
Highlighted
Tourist
13 0 0

Np!

 

I run a data science Shopify app but we don't yet have any "on-site" or embeddable widgets that do this. Have you looked at any existing apps?

0 Likes
Highlighted
New Member
10 0 0

As of yet, Shopify Help has been unable to identify an app providing the service we require. Having a solution designed for us is certainly possible, but there is no assurance it will 'plug-in' with the Shopify platform. This path will likely present an upfront cost structure significantly greater than offer by Shopify. I'm open to any recommendation that will present a more cost effective course. Thank you again for your communication. 

0 Likes
Highlighted
Tourist
13 0 0

Gotcha. Well we have an "off-line" solution where we can do the Customer <---> Product recommendation, which would be much more cost effective as it's a more scalable build instead of a custom project. (https://segments.tresl.co/)

 

If you'd like to chat further though feel free to reach out directly!

0 Likes
Highlighted
New Member
10 0 0

This is my first experience using Shopify, so you have to excuse me if I'm unfamiliar with the capabilities of it's apps. As I understand how a recommendation app works, it recommends items that may go well or compliment a customer's choice. This strategy does not align with our business model. Our strategy is to eliminate choice and filter out the single best product that matches the client profile. I truly appreciate the economic values of using an existing recommendation app, but I'm beginning to think that a costly custom solution is in our future. I would welcome a conversation if you believe it will be productive. Please provide me with your contact details and availability. Best regards!

0 Likes
Highlighted
Tourist
13 0 0

Yep I get it. It sounds like you would want to apply a stricter set of business rules on any product recommendations so the realm of possible products matches a customer's profile. I still think within those products there's room to recommend products, but it will depend on product variation.

 

My email is t@tresl.co if you'd like to chat further

0 Likes