Deciphering Search and Driving Relevant Optimization Part 2

Knowing your retailer is a major step towards understanding how you can drive ecommerce success. When it comes to search, different retailers use different algorithms to determine search ranking. If you can uncover the rules that govern your retailer's algorithms, you can adjust your products to maximize search placement on a retailer site. 

 

Welcome Back

In Deciphering Search and Driving Relevant Optimization Part 1, we covered the challenge that retailer search poses to many businesses and the steps they can take to start decoding retailer search algorithms to drive relevant, retailer-specific, action planning.

In Part 2 of this blog series, we will look at the factors which may or may not influence search positioning. The objective is to rule these factors in or out, depending on the outcome of a trial & error process. These include elements like product availability, promotional activation, and its subsequent clues into sales, and specific product statuses like “New” or “Amazon’s Choice.”

There is a logic to the order in which we approach these further elements. As we go down the page, their influence on search may be more indirect, temporary, or lower than the key pillars covered in Part 1.

Does No Availability = No Visibility?

Availability is the last of the make or break factors. Many retailers will remove your products from consideration in search if it is unavailable for purchase. This heightens the importance of tracking availability and adds weight to any request to your supply chain to remedy the situation. By doing this, retailers are merely protecting their user experience. However, this is not the case with all retailers:

 

ICIPARIS XL

ICIPARISXL Availability
 

The above example can occur in both simple and sophisticated algorithms. The latter may also adapt search behavior against unseen stock levels, to more effectively manage demand for the product online in an attempt to avoid it becoming unavailable altogether.

The telltale sign is seeing a product’s search ranking steadily drop ahead of an out of stock period. A secondary consideration is whether an algorithm then needs to regain confidence in a product’s availability performance before pushing it back up to the prominent positions it once was placed. Therefore, you might see a lag in return to search or a return to lower ranks initially while confidence in supply is restored. If this behavior is noticed, the argument for maintaining good availability for all key SKUs is further heightened.

 

Availability vs. Search Rank

Availability vs. Search Rank

 

Is Search Related to Taxonomy?

While availability can determine product presence or not, so can a product’s positioning in the website’s menu locations if one particular condition is met.

Past research has seen many examples of search results mirroring the ordering of products in a related location within the website’s menu configuration.

Searching for “Koffie” (coffee) in Albert Heijn in the Netherlands, for example, will provide you with the same list of products as the menu location for Koffie.

 

Albert Heijn

Albert Heijn Koffie

 

The extra challenge posed with this sort of practice is trying to identify and test the potential relationships. Coffee is easy to anticipate to a degree, but what if you are a business with a diverse assortment of categories? Also, how do you know whether all search terms are linked this way or just a proportion?

There is one sign that can help you identify whether relationships exist within the search results. One prominent clue, albeit not systematic for all retailers, will be the presence of the browse-node that maps the specific location associated (e.g., Food Cupboard > Confectionery > Chocolate > Dark Chocolate). While this will be identified by manually testing onsite, users of Digital Shelf can compare search and taxonomy results in the Taxonomy Placement feature. Unfortunately, if this browse node doesn’t exist it will require a further test and learn exercise against any search terms of interest.

While this may take a little time, identifying this sort of relationship can shift your strategy entirely. Actions will no longer be made to maximize positioning in search; they’ll be taken to increase your visibility across relevant menu locations.

Can the Retailer Override its Algorithm

While algorithms will have a level of automation to them - that’s the reason for their existence after all - you should assume that they can still be overridden. This may be done in exceptional circumstances or as a regular practice to enable products to be pushed to the top. This is easily observable. A product will simply jump to a prominent rank with no apparent cause, such as sponsoring or promotional activity. 

 

Sephora IT

Sephora IT Exclusive 

 

There will, of course, be a valid reason for it. In the example above, the product is a retailer exclusive, which may not necessarily be a regular factor built into an algorithm. The retailer will have a natural inclination to make the exclusive - a mutual commercial strategy with the manufacturer - a success and will seek to maximize the glance views it gets during the period.

Product forcing such as this can also be used by retailers to plug their own private label products to the top. This strategy is employed by the likes of ASOS* or Douglas**. There is a case to say that margin or profitability might be the factors at play here. Still, if that were true, the products’ ranking fluctuation in question would be regularly interweaving with others. If you see a product pushed to the top and remain there untouched, it will likely have benefited from a manual override.

*British online fashion and cosmetics retailer **European Perfumery

Price & Promotions: Status vs. Sales Impact

With this, you should now be able to determine whether your positioning efforts can be compromised by manual product pushing and whether a simple or more complex relationship exists between search and the availability record.

This elementary vs. intricate question is also one to answer when assessing the potential impact of the last factor in this process. It only relies on what can be seen on the retailer website itself - that final piece requiring internal manufacturer information.

Like availability, product pricing & promotional activity can either directly or indirectly impact a product’s rank due to the increased sales they might temporarily generate. Again the fluctuation behavior noted at the very start of the process will dictate what to look for. Do cheaper or promoted products systematically appear at the top of results for a sample of terms? Ocado***, and by proxy Morrisons*** which runs off the same platform, are two examples where the organic search results are promotion led with no exceptions.

*** UK Grocery

 

Ocado & Morrisons

Ocado & Morrisons Promoted Products

 

This can have two significant impacts on your strategic approach to promotions. The first is to consider this fact if you are due to run supporting media, and more specifically, whether the notion allows you to refocus your investment on areas outside of search if branding (banners) isn’t vital. The second is that you won’t be able to compete for the top search spots all the time, which may result in prioritizing optimization in other retailers and letting your trade plan do the work.

If your product ranking demonstrates high fluctuation levels, then it is likely that sales will be at play within the algorithm. Promotional activity will naturally affect sales here.

You will typically see a ramp-up in search ranking during or after the promotional period as the algorithm adapts to the sales figures it uses to determine the results. The wait will depend on the periodicity of sales data used (daily, weekly, etc.) and whether only confirmed orders are used. This means that product substitutions and cancellations may be considered, resulting in a more pronounced “lag” in the impact on search.

Search Rank and Promo

Search Ranking vs. Promotional Periods

 

This can result in an ecommerce snowball effect whereby a product goes on promotion and drives sales. This drives greater temporary visibility in search going beyond the promotional period, which increases traffic to the product resulting in further non-promotional sales. Add the higher capacity to drive Ratings & Reviews, as mentioned earlier in the blog, and you have a medium-term recipe for success.

When running media you aren't only looking to support a promotion or event. Success is also measured by whether the products in question have greater sales frequency after the campaign has finished. This is precisely what can happen here, so knowing that this can occur within a retailer site is another valuable commodity to leverage, especially when trade planning for online.

Completing the Process and Validating Theories

The final objective of this approach to decoding retailer search is drawing a conclusion for each retailer that stipulates what factors are at play and, if possible, in what order they can affect things - does having the keyword in the title trump everything else for example?

There are a couple of final considerations before doing so, and they rely on your inputs as a manufacturer. These will look into the relative search performance of your known bestsellers, notably whether they are generally better positioned. This will validate that sales have a part to play in search performance, and also whether media campaigns run have had an effect too. This can manifest itself in the same way as promotions above. Still, sites like Amazon also view investment in their media products as favorable and reward products that feature within those investments. These details are not public knowledge, but if you are applying the approach covered in this blog series, then applying this last step can be complementary to ensure all bases are covered.

So, you should now have a far more intricate knowledge of your key retailer algorithms in your market, and the urge will be to go away and start optimizing all products that are key to the areas where your visibility is low. The last thing to do is to test your conclusions before applying wholesale changes. The overall process described in this blog series relies on observation and deduction rather than cold hard numbers and may not be 100% robust.

The best thing is to create a test, using products that are low risk and give you the opportunity to trial individual variables. Optimizing a product title, increasing ratings and running a promotion at the same time may have a tangible positive impact. However, it’s going to be hard to deduce what drove the most significant improvement. Doing one of each across three SKUs would be a more precise method, allowing you to come away with a clear, validated theory to work on your more pivotal products.

 

Three I's of Search

The Three I's of Search

 

Hopefully, through reading this blog series and making use of all of the assets on the Digital Shelf Search Hubpage, you are now armed to go off and begin finding search success in your business. If you are interested in finding out how Digital Shelf can help you succeed in search, or simply help you with your overall ecommerce needs, then please fill out the form below, and we will be in contact with you.

Facebook
LinkedIn
Facebook Share
LinkedIn Share