We use cookies to personalize content, to provide social media features and to analyze our traffic. We also share information about your use of our site with our social media, advertising and analytics partners. For information on how to change your cookie settings, please see our Privacy policy. Otherwise, if you agree to our use of cookies, please continue to use our website.

SEO in the Lab: A conversation with Frédéric Dubut from Bing

Merkle is proud to announce our new podcast, SEO in the Lab, with host Senior SEO Manager, Alexis Sanders. The purpose and goal of the podcast is for listeners to learn about other professionals in the industry, what they are going through, and ultimately garner some advice that can be translated into your career.

In the first episode, Alexis sat down with Frédéric Dubut, Senior Program Manager at Bing.

Below are highlights from their conversation. You can catch the full episode here:

What Should Search Professionals Know about Machine Learning?

Looking at the process of machine learning (ML), it can get complicated from a technical point of view. If you’re not well versed in technical SEO, it can sound like a foreign language. What is important to remember is that ML is a way to generalize search algorithms that are trained with how humans will judge websites according to the programmed guidelines.

The way early models are trained is to have a subset of queries and URLs, and then send human judges to the websites, who are asked to rate the relationships between the query and the URLs according to the search quality guidelines. The ratings are used as validation and as a test set. The test set is used to train your machine learning algorithm. The algorithm should perform well on this small subset of queries and URL’s that have been judged by humans. In the end, thinking, ‘Will my site, according to guidelines, get a perfect or excellent rating?’ is a good way to think about it.

Will machines eventually do 100% of the work? Most likely not. In the end, all of this is built for people. Keeping people involved in the process will ultimately keep the machines honest. Looking at whether the results make sense, and not just that the rating looks good, is most important. Frederic doesn’t think the human touch in ML is going away anytime soon.

What Makes a Strong eCommerce site?

If you look at all the eCommerce websites on the Internet, one question Frédéric says we should ask is, ‘Would we give our credit card numbers to that website?’ Let’s use Amazon’s site as an example. Anyone in the world is confident enough that if they give Amazon their credit card number, they know they will be taken care of and that they’re not going to get unwanted charges. On the other hand, there are many websites that will never get even four digits of my credit card.

Think about the trust factors on a particular website. Does it look professional and have an actual contact address that we can look up somewhere? The idea that a customer is vouching for you is something that you need to take into account if you have an eCommerce website from a trust point of view. The number one thing you want to ask is ‘Are users willing to do business with you, and willing to give you their credit card number?’ This will ultimately impact your presence in search results, since Bing and other search engines are ultimately looking for user satisfaction. It all comes back to trustworthiness.

Frédéric's 3 Key ‘Golden Nuggets’:

  1. Remember that we build sites and products for people. Make sure the products and content we build are useful for users.
  2. Check out Andrew Ng’s Deep learning Course
  3. Sign up for BWT

Want to learn more? Check out the full episode here, or on your favorite streaming platform.

Join the Discussion