A personalization test has produced conversion rates of 40-60% for tailored recommendations, encouraging Kele to build a new customer data platform. The kind of data Kele plans to pack into its customer files illustrates how much more complex personalization is in B2B ecommerce than B2C.

Offering a business customer just the right product when they need it is a sure way to increase sales. But that’s no easy task for a distributor like Kele Companies, which sells 130,000 SKUs and stocks more than 3 million parts at any time, mostly components for the complex systems that control heating, lighting and security in commercial buildings and control systems for industrial facilities.


John Strawn, chief marketing officer, Kele

And it’s gotten more difficult during the pandemic, when supply chain disruptions mean some products inevitably will be out of stock at any time, says John Strawn, Kele’s chief marketing officer.

What’s more, increasingly Kele must make the right recommendations to customers it’s never spoken with, potential buyers who are visiting Kele.com rather than calling a sales rep.

“The shift to digital is prominent,” Strawn says. “While the phone channel has remained consistent, all the major growth has been in the web channel.” He says ecommerce sales grew 30% from 2020 to 2021 and are up another 60% this year.


Kele expects revenue to exceed $250 million this year, with 55% coming from self-service electronic channels. Of that, 75% comes from Kele.com and the rest from custom integrations with large customers’ internal ordering systems that allow buyers to punch out to Kele.com or transmit orders via EDI.

To sustain growth, Kele is building an automated system designed to make the best recommendation to a customer instantly, whether that buyer is on Kele.com, the distributor’s ecommerce site, looking at email or on the phone with a sales representative. To do that, Kele is enriching its customer data, marrying that data with product information and inventory status, and layering on artificial intelligence capabilities that can make automated recommendations.

What we’re doing is to not just turn unknowns to knowns, but turning legacy knowns to really known customers.

proof-of-concept in recent months has produced conversion rates ranging from 40-60% for AI-driven recommendations based on customer behavior data. With that success in hand, Strawn says, Kele is preparing to embark on a larger project that will draw in additional customer data along with information about previous transactions and interactions on the phone and via live chat, email and text.

The aim is to provide even more personalized content by identifying unknown visitors and creating a rich customer file for all customers and prospects. “What we’re doing is to not just turning unknowns to knowns, but turning legacy knowns to really known customers.”


B2B personalization is more complex than B2C

Among the data that Kele plans to incorporate into the expanded customer file is information about the role of the buying group the visitor is part of and that individual’s role within that group.

“They could be a building automation controls contractor on a site trying to estimate a project or looking for the right part,” Strawn says. “Or it may be a project administrator who places the order. It’s a little different experience. One is more informative and one more transactional.”

Understanding the specific role of the website visitor is more important in B2B ecommerce, where many people may be involved in a buying decision, than in B2C, where typically consumers make their selections on their own.

Jessie Johnson Forrester Research

Jessie Johnson, principal analyst, Forrester Research


In B2B ecommerce, “the optimized customer experience isn’t just personal — it’s contextual,” says Jessie Johnson, a principal analyst at Forrester Research Inc. who specializes in digital interactions with customers. “In B2B, that context comes from the added complexity of the decision-making process, where that individual on the website is actually playing a role inside a buying group with members that need to move together through the customer lifecycle, from discovery and evaluation to purchase, renewal, and even advocacy.”

“In addition to product interest, companies like Kele need to know or infer—in the moment—whether that website visitor is an engineer or architect, or contractor or supplier to better understand that individual’s role in the buying group, their information requirements, and when they’re most likely to be engaged.”

Identifying the website visitor is the first step

Kele, which was acquired by investment firm The Stephens Group LLC in 2020, aims to achieve that level of sophistication. But first it had to find a way to track website visitors who do not identify themselves and have not purchased before or created accounts.

Before starting this project early this year, Kele could not identify 80% of the visitors to its website, Strawn says. He says the distributor was already personalizing content for the 20% it could identify, segmenting them into 866 distinct buckets. But it was unable to tailor content for the four in five visitors it did not recognize.


To remedy that, Kele.com began assigning a unique identifier to each unrecognized user, and building what it calls a “customer contact card” for that individual. That’s a file that tracks what the individual searches for and clicks on, how long they spend on the site, how frequently they visit, the product categories and manufacturers that interest them, and more.

If a person registers for an account or purchases, the system captures their name, email address and company. Strawn says the site also pops up offers designed to capture email addresses. That can help Kele identify the company the visitor works for and in some cases identify the individual based on in-house customer information or data from outside sources. All that information goes into the individual’s contact file.

Knowing the branch of a large company where the customer works also helps Kele customize site content and recommendations. One branch may specialize in providing control systems for hospitals,  another for data centers and a third for service industries. Knowing the branch’s specialty enables Kele to fine-tune what it shows each customer.


Identifying the visitor’s sales region also enables the system to add in the name of the Kele salesperson who handles that account, giving that rep more information about the customer. Plus, identifying a visitor’s location can immediately enable Kele to market across channels its promise to deliver within two days to 74% of the U.S.

AI-driven recommendations boost conversion rates

In the testing phase this year, Kele has stored those customer files within Salesforce Personalization (formerly called Interaction Studio) from Salesforce Inc., whose software Kele also uses for customer relationship management and other applications. It’s also added into that software product information, such as which items often are purchased together, and which ones may be delayed by supply chain issues.

As an example, Strawn says, “most of our customers are working a job and need parts fast despite the global supply challenge, so when they view an out-of-stock SKU, we fly a placement that shows an in-stock alternative that can get to them in two days or less.”

The personalization system also makes up-sell and cross-sell recommendations based on the product being purchased, and these recommendations are producing conversion rates well above the strong 4.25% average for Kele.com. When the system recommends additional products typically required with the purchased item, the conversion rate is 62%, and it’s 42% for related products, Strawn says.


The role of a customer data platform in B2B ecommerce

That success has convinced Kele to move forward with a more ambitious project that will allow it to incorporate more information about each customer and to provide recommendations to sales reps and anyone else who interacts with that individual.

Kele is currently with vendors of customer data platform software that would house first-, second-, and third-party customer information that can be paired with cross-channel intent signals. Kele will use this CDP to recommend the best product or action for the customer to take next, based on the channel they’re interacting with. Strawn says he plans to have the new application up and running by the second quarter of 2023.

Where the customer files in Salesforce Personalization store about 100 attributes of each individual, the CDP will be far more robust, enabling Kele to collect considerably more data points about each individual, Strawn says.

It will also enable Kele to collect information about recent interactions, such as if a customer raised an issue with a delivery in a recent phone call or on-site live chat. In that case, Strawn says, a scheduled marketing email may be replaced with one that offers the customer help in resolving the current problem.


Once the personalization system is operational for the building-automation side of Kele, Strawn says it will be extended to industrial control systems. In recent years, Kele has made two acquisitions of companies that provide industrial controls.

Enabling sales reps to help customers more effectively

The overall aim, Strawn says, is to help customers find just what they need as quickly as possible. One way Kele plans to do that is to add a click-to-call button to Kele.com so that a customer struggling to find something can be immediately connected to their Kele sales rep, who will have access to the customer’s data and be able to see where the person is on the website.

“We’re looking at more real-time integration with our phone system so the customer can click to call, and the rep can see where they are in the session while the AI and machine learning spits out recommendations,” he says. “The rep can lead with that, as opposed to spending five minutes figuring out where they’re at and what the problem is.”

The goal, Strawn says, is to enable Kele’s sales reps to provide a more real-time, relevant recommendation to the customer, no matter what channel they start and finish in.


“This is about creating a customer-centric experience that is highly personalized based on all the data we have available to us and passing the next best action or offer into whatever channel the customer prefers to use,” Strawn says.

“That could be Kele.com, so we will surface personalized recommendations at key points in their shopping experience. It could be calling into our sales team, so we will pass that data to the sales rep so they can have an informed conversation with the customer and get them what they want faster, or it could be an email if they abandon the shopping experience in order to pull them back into the funnel.”

Business buyers increasingly expect such intelligent recommendations and digital content tailored to their interests, says Forrester’s Johnson.

“From their experiences as consumers, B2B buyers and customers are already accustomed to receiving relevant content and product recommendations curated by algorithms and delivered at the right time and in their preferred channels and formats,” she says. “B2B organizations are finding new ways to leverage AI and automation to better understand, engage, and enable their buyers through more personalized and connected interactions.”


Kele already is seeing the benefits of its personalization efforts in higher conversion rates. And it hopes to see even better results across more channels as it builds a system with more data about its interactions with customers and prospects.

Sign up

Sign up for a complimentary subscription to Digital Commerce 360 B2B News, published 4x/week, covering technology and business trends in the growing B2B ecommerce industry. Contact Don Davis, editor at large, at [email protected] and follow him on Twitter @DonDavisDC360.com.

Follow us on LinkedIn and be the first to know when new Digital Commerce 360 B2B News content is published.