Lima Vallantin
Lima Vallantin
Data scientist, Master's student, and interested in everything concerning Data, Natural Language Processing, and modern web.

Contents

Don't forget to share:

Share on linkedin
Share on twitter
Share on facebook

Don't forget to share:

Share on linkedin
Share on twitter
Share on whatsapp
Share on facebook

For the past months, I became more and more interested in how machine learning and automation contribute to a change of landscape in the SEO and digital marketing field.

This is a theme that passionates me because I know how boring and repetitive some of these tasks can be. I also know that a small pinch of Python and a basic knowledge of machine learning may increase our work speed.

Let’s take the Mad Affinity example. Mad Affinity is an art print e-commerce where I have been applying some automation logic to improve our social media reach and product catalog management.

We need to rework each new product published on the platform. This includes:

  • Analyzing the colors’ pixels on the image to create a color scheme.
  • Create supplementary color schemes for the first color palette.
  • Defining a couple of keywords for the product, that will be linked to the website’s internal search.
  • Writing a short and a long description.
  • Adding a relevant small text talking about the theme of the art print to increase the text/HTML ratio.
  • Looking for suitable cross-sells on the website.
  • Adding the images.
  • Translating the content.
  • Creating a small text for social media.

As you can see, each one of these tasks take time. Just looking for the colors and the colors names is extremelly time consuming.

Each product has its own color scheme. Manual creation of each scheme would take a lot of time.

Somebody call python, please

It’s no secret to anyone how I love python. This programming language is platform agnostic and is a great pick when talking about automation.

We discussed a lot about killing some of our product’s page features in name of agility, but we came to the conclusion that these features offer value to the customer.

The solution to continue to offer these features without losing agility was to use python.

For the first part of the product creation, we use a clustering algorithm to check the pixels colors on the picture and to determine the main ones. Then, we send these colors via API to a color server, which returns the name of the color and its complementary scheme.

The next step is to build a description text using a text spinner and to generate a second text with the main keywords for the product.

Another step is adding cross sells to each product. We also automated this part, simply using Selenium. We then add the images and send the description also via API to the online website, where a final revision is made to be sure that everything makes sense.

The script continues to run to produce a social media descriptive text for the product.

Without automation, doing all this could take forever. Using python removes the need of appealing to external paid tools, what also saves us some money to invest on another things.

I can’t stress the importance to new and old marketers to learn how to program, or at least, to learn how to use basic free tools that we can find for free on GitHub.

It’s time to end the thinking that when it comes to automation in marketing, we are only referring to sending emails.

And you, as a marketer, how could you use more automation? Do you see new opportunities for using automation in marketing?

Don't forget to share:

Share on linkedin
Share on twitter
Share on whatsapp
Share on facebook

Leave a Reply