Recap: Understanding AI and How it Works
In “Understanding AI and How it Works,” the first webinar in our 4-part NAW webinar series, we looked at how customer expectations are higher than ever, even for B2B buyers. This has largely been driven by the rise of technology-first companies like Amazon and the shift to a more digital-dominant sales environment due to COVID-19. These heightened expectations accelerated the need for distributors to adopt AI in order to become more efficient and effective, enhancing the customer experience across all touchpoints.
AI is not like other distribution technologies. Benefits from investments in PIM, sales order automation, and punchout catalogs, for example, scale linearly — no matter how many additional customers you onboard to your punchout catalog, the cost savings per customer remain unchanged. AI-first solutions, on the other hand, scale exponentially. More customers and more sales drive an exponential rate of growth in the AI’s performance.
Behind that growth is data, the foundation of AI. AI solutions need very large data sets to improve the accuracy with which they work and the predictions they make. Distributors have those very large volumes of data, generated by years of transaction history along with data from CRM, ERP, ecommerce web sites, product catalogs, product images, and more. The data may be noisy, but the predictions generated by deep learning models can reach high levels of accuracy despite the noise.
In case you missed it, you can watch the webinar replay. For a deeper dive, check out our whitepaper on the same topic. Read on to learn how AI enables distributors to deliver the personalized experiences that B2B customers value.
Understanding AI Personalization
Personalization is AI’s ability to deliver an individualized experience for every customer in real time, rather than grouping people into broad categories or market segments. Before segmentation or personalization, companies would market a selection of best-selling items to everyone. They might also key off of commonly experienced parameters like events or seasons — you don’t sell snowblowers in July — but the sales and marketing were essentially product-driven, not buyer-focused.
Advanced marketers began to manually break customers into segments to make more targeted pitches. This practice is still common in both the consumer and B2B space. Consumer segmentation, for example, is often based on demographics like age, income bracket, or location. Similarly, B2B buyer segmentation is based on firmographics, such as industry, company size, and location.
The common thread is that manual segmentation requires humans to dig into the data and attempt to define common groupings of individuals or accounts. For B2B marketers, this could take the form of loading customer data from a CRM system into a spreadsheet and sorting by state or ZIP code. That way, at least, you don’t try to sell snowblowers to companies in Florida, even if it is winter.
Even though segmentation is an attempt at customer-centric marketing, the groupings are still broad, with no assurance that individuals or accounts falling into those groups do indeed have the need or desire to purchase the product(s) being marketed and sold to them. All too often, the offers missed the mark because segmentation is not specific enough.
With AI, personalization targets individual customers, not groups
Personalization goes beyond segmentation. It comes from understanding an individual’s actions, and the context in which those actions take place, not just from knowing what that person or account “looks like.” For B2B companies, those actions can include browsing patterns, purchase history, responses to marketing “touches,” and more— did the customer open that email? Read that white paper? Buy that product?
AI analyzes this data to create a customer profile, updating that profile dynamically by learning and changing as new data is gathered. This enables sales and marketing to predict the best item for each customer, meeting the ultimate goal of presenting the right product or service to the right person, at the right time.
While humans could hypothetically dig into an individual’s actions to create a personalized experience, imagine what it would take to manage that for thousands of people with potentially dozens of interactions each. Furthermore, this effort would be subject to human error. AI’s ability to analyze large swaths of data in a short period of
time empowers businesses with precise targeting at scale.
Three Ways AI-driven Personalization is Used
True personalization is helping brands be more competitive and is changing the face of sales in the B2C space. This same change is actively taking place in B2B distribution.
Here are three ways that AI personalization can work for distributors.
Content personalization It’s incorrect to assume that B2B buyers don’t read content online when researching a purchase. According to Gartner, business customers visit a supplier’s website frequently over the course of the buyer journey, and 70% of that research is done before the supplier is selected. Those are prospects you want to win over.
AI models can predict what type of content a returning site visitor is most likely to engage with based on that visitor’s historical content consumption. This affords the opportunity to accurately target content to an individual and offer higher-value content that encourages them to buy from your brand instead of your competitors.
Tailored omnichannel customer experience McKinsey reports that B2B omnichannel adoption exploded during the pandemic, condensing a decade’s worth of digital transformation into months. Their survey of 3,600 B2B decision makers found that during the pandemic more than 90% of B2B companies shifted to a virtual sales model. Among customers, 70% prefer remote human interactions or digital self-service, and 97% of that group are willing to make digital, self-serve purchases exceeding $50,000.
AI personalization enables distributors to provide continuity and an ideal customer experience across all channels, including online and in person. Without AI, channels are siloed and unoptimized. In advanced implementations, AI systems track all touchpoints for each individual. So, when a qualified prospect moves from digital to talking with a sales representative, that rep will know the history of the “online conversation” and pick up where that left off.
Upsell and cross-sell opportunity identification Seasoned salespeople who know your catalog inside and out, or at least know where to look, are good at upselling and cross-selling. But increasingly, that deep domain knowledge is hard to come by. AI models can codify that knowledge and identify products that a customer is likely to want or need based on data like purchase history, product categories and your current inventory. These recommendations can be served up online, either on a product page or at checkout. They can also be integrated into sales workflows so the insights surface when they can help reps sell additional products.
Netflix’s Content Personalization
When we think of AI personalization leaders, the Netflix streaming video service should be top of mind as having the most successful recommendation engine. Each Netflix customer gets personal recommendations about what to watch next, based on viewing history, and every time a viewer clicks on play, the algorithm updates. The more you watch, the more up-to-date and accurate the algorithm is. It’s so accurate that 80% of Netflix viewer activity is driven by personalized recommendations.
Netflix also tracks more than what you watch. It tracks when and where you watch, what’s in your user profile, the device you watch on, your search history, your browsing history, whether you pause or fast forward, and whether or not you watch the whole movie, episode or series, and more. It’s all done in the pursuit of a user experience that will keep you engaged with Netflix, such as serving up a homepage that is totally personalized for you.
Netflix is a data-first company The massive amounts of data that Netflix processes also drives its original content and content marketing. Unlike most content creators, who rely on experience, intuition and luck, Netflix uses data to decide what original shows and series to produce. Which has the best track record? The typical TV show has a 35% chance of succeeding. Netflix original content is successful 93% of the time.
Think of Netflix shows like “Queen’s Gambit.” You wouldn’t think that a show about a chess master would be wildly popular, but this one was, setting viewership records at its release. The company also uses data to determine what promotions you see for their original content, creating up to 10 promos and multiple thumbnail images for homepages that are served up based on what Netflix knows about your preferences.
Netflix’s personalized customer experience and AI-driven content strategy has resulted in significant subscriber growth. As of Q2 2021, Netflix has about 209 million global paid memberships, which grew to a record 213 million in Q3 2021. With more subscribers, Netflix can better determine what content people like, use that insight to create content that people love, and recommend the right content to each viewer.
Home Depot and the Omnichannel Experience
You wouldn’t necessarily think of “hardware” and “digital creativity” in the same sentence, but Home Depot has made that combination into a winning omnichannel experience for its customers.
According to CEO and president Craig Menear, the retailer’s success during the COVID-19 pandemic is a result of its omnichannel efforts and long-term investments in technology.
The company’s mobile app sits at the center of its strategy, providing a consistent and continuous experience whether the customer is shopping online, in the store, or both. In 2020, half of the online orders were picked up at a Home Depot store, and the website continues to be a growth engine for the business. Behind the scenes, the company is employing AI to enhance the customer experience. Here’s how.
Capturing customer data to personalize shopping By capturing and consolidating customer data from activity on the mobile app, on the website and at the store, Home Depot is able to create a comprehensive database that captures customer personalities and buying behaviors. Using data on shopping behavior, product affinity, and other insights, the retailer can tailor an individual shopper’s experience and realize higher conversion rates on product recommendations that, with AI, get better over time. Home Depot’s AI can also identify which items are similar to one another and recommend substitutes if an item is out of stock.
Using customer data across channels Home Depot’s omnichannel strategy also includes using data from one channel to power AI analytics that drive improvements in other channels. For example, data from the website and mobile apps can reveal which are the best-selling products overall. This insight is then used to determine how much of those products to stock in retail locations and where to put them in the store. The same data can be used to prioritize how items are listed on the mobile app and website.
Image and voice searching AI image recognition models enable customers to take a picture of an item in order to get the information they need. The app can tell them what the exact product is and link them to the product online or identify which nearby retail locations have the item in stock. The mobile app also uses AI voice technology to enhance search. Customers talk to the app like they would a smart assistant, and the app will search through more than 1 million items to deliver relevant results.
Understanding what customers want to use products for Based on the products a customer searched for and purchased, Home Depot’s AI can understand what type of project the shopper is working on. The app then serves up project how-to and buying guides to help customers get all the knowledge and products they need to get the job done successfully.
Amazon’s AI-Powered Strategies to Increase Sales
To call Amazon a sales juggernaut may be an understatement. The company had third quarter 2021 sales of over $110 billion, an increase of 15% over the previous year. Of greater interest to distributors is the progress of direct-threat Amazon Business, which the company says has reached $25 billion in worldwide annual sales.
AI is at the center of the company’s success. It’s AI-powered approach, which utilizes product recommendation and advanced search, helps them dominate markets. This technology has delivered huge results, with AI now contributing 35% of their online sales. Distributors need to understand and adopt these same AI-powered strategies to fend off competition, delight customers and fuel business growth.
Product recommendations One of Amazon’s most effective, and most copied, strategies is using AI to process customer data and make personalized product recommendations. Bounce rates – the percentage of customers that do not click beyond the home page – tell the story: Amazon’s rate is 35%, compared to 45% for Target and 50% for Walmart. AI recommendations move customers through the site and nudge them towards sales. With a conversion rate of 13% — compared to 2% for the average website — it’s no surprise that in 2020, $69 billion of $197 billion in online sales came from AI recommendations.
Here’s how it works. As customers progress through the shopping experience, Amazon predicts what items each individual is most likely to buy. With each new page, customers see a fresh batch of personalized product recommendations to further increase spending. Product pages feature the main product along with personalized recommendations like “recommended for you,” “products you might like,” “frequently bought together,” or “customers also bought.” The product recommendation strategy carries through to checkout when customers get prompted with promotional sections labeled “recently viewed” and “saved for later.”
Amazon’s recommendations strategy measurably increases customer engagement with the website. SimilarWeb.com reports that average visit duration for Amazon is 7:24 minutes, compared to 5:24 minutes for Walmart and 4:54 for Target. Amazon visitors click on an average of 8.9 pages during their visits, while Walmart gets 5.3 page visits and Target, 5.4.
Advanced search relevancy When a customer uses the Amazon search bar, they are basically declaring their interest in a product. It is natural to think that search doesn’t matter very much, and that when a customer starts looking for an item on a website, they are probably going to end up buying it no matter what.
However, the data clearly reveals that this does matter. In fact, it is worth more than $10 billion to Amazon. When you search for something on Amazon, you are four times more likely to buy it than when searching on comparable sites. In fact, Amazon enjoys an amazing 12.9% conversion rate when customers use their search bar. Rivals like Walmart and Best Buy, in contrast, convert only 2 to 3% of customer searches into sales.
Why is Amazon search so good? It uses AI to understand a customer’s search and its context to return more relevant results to users. A shopper searching for a brand name that Amazon doesn’t sell will be presented with close matches. Even if the customer searches for a brand name that is not sold on Amazon, the results will also show best-selling and sponsored items as well as results for other brands — delivering relevance and choice.
Relevance and choice, in turn, translates into clicks and sales. One study showed that 42% of Amazon searches result in a click. The competition fares poorly in comparison: 16% clicks for Walmart, 13% for Target and 12% for Etsy. If Amazon search performed like everyone else, they would miss out on nearly $800 million a month.
Data Network Effect
When data lives on a platform like Netflix, Home Depot, or Amazon, the platform can capitalize on the data network effect. Forbes describes it like this: “as more users join a platform, more information, data products and content are produced — all leading to increased innovation and market value for the platform.”
Another way of defining the data network effect is that it creates a virtual circle. More users leads to more data, more data leads to better analytics, better analytics leads to a better product, better products lead to more users. This is a sister concept to the AI-powered digital operating model, which we discussed in our first white paper in this series. AI-and data-driven companies can scale exponentially by connecting with other digital businesses and their data to create new opportunities and growth.
Why B2B and B2C Recommendations are Different
While offering online upsell and cross-sell recommendations may look like adopting a B2C strategy, it’s not that simple for distributors. B2C product recommendation engines — AI systems that analyze data and push out suggestions for related products — don’t work the same way for B2B. That’s because B2C engines primarily rely on online activity, or what people are browsing for, to select and display product recommendations.
Making useful and accurate recommendations for B2B buyers requires using historical transaction data for that customer, product attribute data, other customer data, and data from all the different channels that distributors sell through. These include the distributor’s own ecommerce sites, marketplaces, punchouts, PIM, EDI, and people (sales reps in the field or on the phone, plus customer service reps).
Product attribute data is also more complex, because distributors have far more SKUs than most B2C businesses. Distributors need a product recommendation engine that analyzes all that data to provide relevant and useful cross-sell or upsell opportunities to either their online customers or their sales reps.
Another difference between B2B and B2C to take into consideration is customer purchasing behavior. Consumers are typically making one-off purchases and can be tempted by recommendations to make an additional impulse purchase. B2B buyers are more likely to make regular reorders or buy items for specific projects. They might also be searching for what’s new or for opportunities to broaden the products they carry. Time frames tend to be urgent, so if you are out of stock, your website or sellers need to offer a workable substitute from inventory.
A final difference to note is that B2B buyers still prefer to talk to a person. According to McKinsey, only about 4% of buyers buying on behalf of a business prefer digital for all buying scenarios, and 76% want to talk to a person when considering a new product or service.
Once we understand the differences between B2C and B2B recommendation engines, we can look to the consumer space for inspiration and to see just how much of an advantage AI-powered personalization can give to companies that engrain it into their business.
Distributors have significant opportunities to replace manual customer segmentation with AI-driven personalization to deliver a consistent, tailored customer experience and provide upsell and cross-sell recommendations that will boost the bottom line.
Category leaders in the B2C space are the companies that are investing in AI and novel use cases, including Netflix for streaming video, Home Depot for retail, and Amazon for ecommerce. Although distributors have more complex, multi-channel sales structures and very different buyer profiles, they also have much more data to work with than comparably sized B2C companies.
Taking these differences into account, distributors can win with AI by adapting B2C AI strategies in 4 ways:
• Deliver a consistent omnichannel customer experience. Take data out of channel silos to provide a seamless experience across touchpoints for customers and keep sales and customer service reps informed about all interactions. Use AI analytics to drive improvements across all channels
• Develop a content strategy based on personalization. Use AI to understand what content your customers want to engage with as they research new products and plan projects, and serve it up at the right time and place.
• Structure AI recommendations for B2B sales. Consumer websites nudge shoppers toward impulse buys, but that tactic doesn’t work the same way for distributors. Instead, B2B sellers must help customers fill out regular re-orders, make complete project orders, and find good upsells, cross-sells, and add-ons.
• Invest in AI-powered search. The primary impact of using AI to enhance search relevance will be increased ecommerce revenue for distributors — and happier customers. Distributors can also use that technology to help counter sales reps locate products and substitutes and make it easy for telesales and field sales to find the right products for their customers.
The clock is ticking on the advantages you can realize with AI. Early adoption can be the difference between 122% and 10% gains. In order to win, distributors must beat competitors to the market.
About Proton.ai Proton.ai was founded in 2018 by Benj Cohen, fourth-generation distributor and Harvard alumnus. Proton is an AI-powered sales enablement platform, purpose built to increase revenue for distributors by helping sales reps and customers navigate the complexities of managing lots of products through multiple channels. Proton helps distributors grow revenue by 5%-10%+ and gain market share.
About NAW The National Association of Wholesaler-Distributors (NAW) is composed of direct member companies and a federation of international, national, regional, state and local associations and their member companies, which collectively total more than 30,000 employers, with locations in all 50 states and the District of Columbia. NAW-affiliated companies are a constituency at the core of our economy—the link in the marketing chain between manufacturers and retailers, and commercial, institutional and governmental end users. Industry firms vary widely in size, employ more than 5.9 million American workers and account for $5.3 trillion in annual U.S. economic activity.