5 considerations for Black Friday 2023 readiness

It’s hard to imagine having to think about Black Friday less than 4 months out from the previous one, but that’s exactly what many retailers and ecommerce companies are doing according to the recent Ecommerce Search Trends 2023 report. Many have already begun to look for innovative ways to scale and pivot in preparation for the upcoming holiday season.

Black Friday automation and scale

Black Friday puts huge demands on the entire organization. It has also become a bit of a moving target. In 2022, consumers and retailers got an early start on Black Friday. Deals were going up more than a month in advance. Spending was driven by deep discounts, something that consumers waited for after skipping past the early-holiday deal days from major retailers such as Amazon and Target. 

Holiday discounts and markdowns played a key role in the year-over-year rise in December sales, but they weren’t enticing enough for consumers to show improvement from November. But, that didn’t mean everyone was buying right away. Contrary to the height of the pandemic in 2021, where consumers were willing to pay more for goods during peak season for fear that retailers will run out of inventory, consumers in 2022 were well aware of the inventory overstock and decided to hold on to their wallets while waiting for the latest and greatest deals.

With these kinds of dynamics — campaigns that may need to be published and changed quickly to respond to competitors, buying trends, and inventory feasts and famines — are putting pressure on companies to find solutions that rise to the challenge. Survey respondents said they have already begun to look for ways to automate labor intensive workflows, help them scale, and iterate on the go.

Black Friday technology considerations

Given all of the complexity around Black Friday preparation, we have compiled some recommendations worth consideration. 

• Maintaining consistency 
• Blending manual curation and AI 
• Fat head, chunky middle, and long tail optimization
• Instant indexing for accuracy 
• Performance and scalability for uptime

Maintaining consistency

Holiday sales are not only fast and furious, but dynamic and changing. It could be very easy to remember to update one thing while forgetting another. In addition to finding digital asset management solutions for managing creative, merchants will want to consider where and how information is being shared. 

Whether someone is using your search bar or browsing product listing pages (PLPs), Black Friday merchandising needs to be consistent everywhere to maximize sale opportunities. This is one area that search technology can help. Many PLPs are entirely dynamic — they are, in effect, queries — just like on-site search results. In other words, the systems you have in place should be able to push Black Friday sale items to the top of search results and PLPs. It should bring a measure of relief knowing that the content you prepare is delivered everywhere your brand buyers are shopping. 

Oh, and don’t forget mobile! Mobile search and PLP design and delivery need to be responsive to maximize sale opportunities for buyers on the go or on the couch! 

Blending manual curation and  AI 

AI performs really well for specific use cases. Machine learning algorithms can be very good at prediction and personalization. It can greatly augment the work your team is doing and calculates results at lightning speeds to deliver the right offer to the right person at the right time. Nonetheless, there are times when you will want to take the reins. 

For example, you might want to direct your merchandising solution to apply AI-based optimizations on specific parts of your catalog, but manually curate other areas. AI is great for boosting results based on real-time data, such as pushing best-selling or high margin items to the top. On the other hand, rules are great to create immersive brand experiences and control sponsored items. So, when evaluating AI solutions, enquire about control and flexibility.

Fat head, chunky middle, and long tail optimization

‘Tis the season for your buyers to search your site for “the best scarf to give nana” and other long tail and infrequent queries. These long tail queries can make up 50-70% of the queries on your site. They’re incredibly difficult to predict, and there’s no easy way to write rules and synonyms for all of them (like “nana” is synonymous with “grandma” etc.).

More broadly, people are increasingly searching like how they think, and therefore a search solution should be able to decipher query intent to return relevant results.

When it comes to scaling for Black Friday, a search solution needs to handle these many types of queries — and this is where end-to-end AI comes into play. 

• Query understanding: A search engine will need to parse the query itself. AI query understanding models such as natural language processing (NLP) will use features such as sentence encoding, typo tolerance, and lemmatization to prepare the query for retrieval.
• Retrieval: AI retrieval is where vector search comes into play. Vectors are the technology behind AI understanding, and they allow visitors to search as they think. People can type “red sparkly dresses” and get sequin dresses even if the phrase “red sparkly” isn’t on your website, for example. Vectors are good, but for scale, you’ll need it to have neural hashing and keywords, too. The combination is sometimes called hybrid search, and it’s the only way to optimize a wide spectrum of  queries — from the fat head to the long tail.
• Ranking: AI ranking capabilities such as reinforcement learning will continually push better results to the top. “Better results” means results that are getting the most clicks, conversions, purchases, ratings, or other positive signals.

For Black Friday scale, AI “hybrid” search can improve discovery for all three kinds of search — from the fat head to the long tail. AI search not only displays the most relevant results but also expands the footprint of product discovery. Hey, and who wouldn’t want to make this all easier for the consumer to just ask Alexa, Google, or Siri to ‘add a nice scarf for nana to my shopping list’ and have it in my [brand name here] basket?

Instant indexing

Many retailers and marketplaces update their search index nightly. Sometimes hourly. Updates are crucial for reflecting the latest product varieties, prices, new SKUs, inventory levels, and more. Your search provider needs to be able to handle frequent and massive changes, which can also include curveballs such as schema update, customer reviews, new merchants, etc., and it needs to make these changes without affecting the search performance. When selecting a search provider, it’s good to test the indexing capabilities to be sure they’re up to the job around the holidays – this is a true test of speed at scale. Remember, if your buyer can’t find it, or finds things that are not available, that doesn’t lead to good outcomes.

Performance and scalability for uptime

Black Friday puts massive constraints on systems — all your systems. Many of them either can’t scale or require in-humane spending to be prepared. Think about architecture that allows the business to scale without 10x increase in your spend to adjust systems. Modern cloud architectures allow you to not have to manually ramp up and down dedicated infrastructure.

Network latency creates real problems for retailers at Black Friday, so it is imperative to identify ways to get faster response times with modern cloud-based architectures across all services on your site. For example, API-based Search platforms such as Algolia are an effective way to limit the amount of traffic to your commerce and content platforms for browse and discovery, and not just search use cases but also when buyers are browsing PLPs. 

Black Friday may feel a long way off, but it has a surprising and sneaky way of arriving soon. As you gear up for the big sales event, it’s important to find solutions to help you scale and maximize margins. Download a copy of the 2023 Ecommerce Search Trends report, or contact us for more information about how Algolia might be able to help.