A PERSONALIZATION RUCKUS – MARKETING TODAY, TO DIGITAL SHELF PERSONAS TOMORROW

A PERSONALIZATION RUCKUS – MARKETING TODAY, TO DIGITAL SHELF PERSONAS TOMORROW

In part two of my ruckus musings, I focus on personalization’s impact on measurement, but more importantly, the future of the digital shelf. Make sure to go back and read my retail media measurement ruckus musings in my past newsletter.

Not so long ago, I was a market researcher. I have run multiple segmentation studies throughout my career. Segmentation was always about unlocking different personas, usually, a maximum of 7 personas, to appropriately cater products and messages to those ideal targets. In one study I had a segment that was a dad who liked to make Christmas magical for the kids, so he would buy the hottest brands. How we messaged and targeted him was different from others.

As time progressed, segmentation became about looking at the path-to-purchase and customer journey. This too was persona-based, such as the busy mom journey. The busy mom's journey, like other journey paths, was not linear or universal. This led to micro-segments, a segment of a persona. One busy mom may really enjoy salty snacks as a treat and enjoy dancing, while another busy mom may be gluten-free, buy salty snacks only for the kids, and would rather read books.

The power to personalize and re-target has been a long-standing trend in eCommerce. Even today, you see the item you searched for five minutes ago - or just talked about to your partner - everywhere.

The Tom Fishburne cartoon shown below demonstrates the interconnection between online shopping personas and in-store shopping. This is the holy grail for retailers, a trend that was evident at Groceryshop last year. Look closely at the cartoon’s bottom left corner… it was created in 2011! The camcorders, MP3, and Facebook references may be the biggest clues. I know you were all excited to see MP3s return. I didn’t get a smartphone until late 2012.

No alt text provided for this image

With the continued push toward a first-party data future and programmatic audience targeting capabilities, these micro-segments are attainable in media and marketing today. Micro-segments will be tied to automated solutions that feed personalization across every retail touchpoint a consumer interacts with. As noted by the CEO of Upside at Groceryshop last year, in the future retailers will be looking at margin consumer by consumer; with personalization unlocking different value and margin plays.

With this, a retailer could curate their online shopping experience specifically around an audience who bought Cheetos, Mountain Dew, and viewed protein drinks, but didn’t buy them. You may think it indicates a male Gen Z, but if you also found out they bought dog food and women's shaving cream, you might think differently.

As retailers are getting sophisticated with their first-party retail media audiences, for general online shopping, most personalization of U.S. online shopping sites by retailers is still primarily around targeting recommendations, recipes, and order history. Because of privacy laws, all of these are difficult to measure using digital shelf web scraping.

Currently, online shopping organic and paid search results are largely keyword-based or influenced by algorithms around content and sales performance. However, this will continue to shift. My friend Chat GPT provided a clear context for how retailers personalize the digital shelf. [I’m still on the wait list for the new Bing search.]

Collect data: To personalize the digital shelf, you need to collect data on your shoppers' behaviors and preferences. This can include information such as their purchase history, search history, clicks, and views.

Segment your audience: Once you have collected data, you need to segment your audience based on common attributes such as age, gender, location, purchase history, and search history. This will help you to group your audience into specific segments and provide a more targeted experience.

Use personalization algorithms: Personalization algorithms use the data you have collected to provide recommendations to shoppers based on their past behavior. These algorithms can use machine learning to analyze the data and provide more accurate recommendations over time.

Test and refine: Personalization is an ongoing process, and it's important to continually test and refine your personalization strategies. This can involve A/B testing different product recommendations or experimenting with different data sets.

Personalization causes conflict with web scraping, the primary input to digital shelf measurement. Depending on the keyword, my share of searches may be acceptable. However, did your keyword bid account for the busy mom who is gluten-free and buys salty snacks only for her kids, or the dad that wants magic for his kids? And second, how do you measure your true digital shelf presence when this is all privacy locked in web scraping? It may come down to those personas.

However, in places like China, where the search results are personalized, the true search results and audiences are housed in walled gardens such as Alibaba.com's Brand Databank which absorbs and deposits every online and offline interaction between brands and consumers, and tracks the status of brand consumers throughout the awareness-interest-purchase-loyalty and are activated across Alibaba's ecosystem. Known web scraping companies in China can only scrape blank accounts without history, which one person told me was a neutral setting to assess for new users.

Uber Eats is understanding its Eater's intentions in its restaurant personalization technology and is now applying it to its growing grocery business. For restaurants, UberEats considers restaurant popularity, restaurant attributes, contextual factors such as dinner or weekday traffic, and Eater (user) behavioral attributes such as order and cuisine frequency, along with their clicks and ratings of restaurants.

No alt text provided for this image
https://www.uber.com/blog/uber-eats-recommending-marketplace/

In the end, consumer brand companies need to combine walled measurements with neutral web scraping. The way to do this may be the same as retailers: collect data, segment your audience, use personalization algorithms, and test and refine.

This means that AI, and Generative AI, will become more relevant for both brands and retailers to train digital shelf measurement for personalization. Targeting the digital shelf in addition to the onsite ad audience. This will include layered keywords and different segments/personas for digital shelf images, product titles, and PDP product details/images.

The digital shelf may become 7+ digital shelves to manage and could become programmatically targeted. Testing and learning is the only way to learn which personas matter for the retailer and your brand.

Visual attention measurement has been growing for programmatic digital marketing and testing digital shelf images, which uses AI trained in neuroscience, across audiences, placements, and content. In a past role, I was trained in neuroscience and did a lot of neuroscience-based ad testing.

Using EEG, an electrical activity measurement of the brain, neuroscience has the ability to understand subconscious audience engagement that is triggered by personal relevance. AI that assesses neuroscience-based engagement along with visual attention will help curate to different audiences. However, we still are curating digital shelf content mainly to one or two target audiences or basic product features, and not a series of personas or demos to maximize conversion.

To understand these differences further, I once again referred to my friend ChatGPT and asked about product tags for two demographics for Cheetos, Male Gen Z, and Middle-Age Women. See responses below.

No alt text provided for this image

Cheetos' core target audience is usually thought of as youth-oriented. At first look, Cheetos.com appears to target Gen Z or Generation Alpha, but some recent ads target families and couples. So instead of only targeting my ads, perhaps I should target my digital shelf. Using ChatGPT’s insights, I changed the hero image banners on Cheetos results on Walmart.com. While this would need to be tested, could a more personalized tag work? Then you would need to take into account the other digital shelf elements.

No alt text provided for this image
My augmentation of Walmart.com Cheetos search results and their hero image banners

We tend to pick one primary attribute to highlight across the content focused on functional product attributes, which do have merit. Brands also face bandwidth constraints and limitations to update content. Personalization would make content updates more cumbersome.

OpenAI's ChatGPT and Dall-E, however, teach us that Generative AI can automatically generate images and copy. Coca-Cola made this future clear this week with an announced OpenAI and Bain & Company partnership to explore use cases for personalized ad copy and targeted messages.

As I mentioned earlier, personalization was a hot topic at Groceryshop last fall. Albertson’s Chief Merchandise Officer showcased a video looking at two types of shopper motivations, one looking for ease and one for inspiration.

With this push by retailers, it is not just media and marketing that should play along. It should also include the digital shelf, which will change the way we look at Share of Search in the future. You just need to understand your layers of personas, how they may be impacted by behavioral targeting in search results, and test, test, test…and learn.

#ecommerce #retail #digitalshelf #commerce #personalization #ai #chatgpt

Barry Thomas

Global Customer & Marketing Leader | Digital Commerce & Marketing Practitioner | Hospitality Growth | Future of Commerce | Board Advisor | Entrepreneur

1 年

Outstanding insight and pov my friend!

Mike Black

Chief Growth Officer at Profitero. eCommerce Geek.

1 年

Celia Van Wickel Great article. For content personalization to scale I think it needs to be built into the retailer website technology vs something that brands generate in their own walls and then pass to the retailer. Right now brands who create seasonal specific content struggle to get it to the retailer on time to matter so you can only imaging how slow the deployment and upkeep of personalized content will be. Perhaps retailers should see personalized content as a monetizable service similar to Retail Media but instead of using the retailer 1P data for ad targeting you are using it for content targeting? In short, either the retailer or a content syndication company / DAM has to be in on the game. The problem is not generation but deployment.

CHESTER SWANSON SR.

Next Trend Realty LLC./wwwHar.com/Chester-Swanson/agent_cbswan

1 年

I'll keep this in mind.

要查看或添加评论,请登录

社区洞察

其他会员也浏览了