As the senior marketing executive for four publicly held distributors, I’ve led thousands of product promotions. Generally, for a large distribution company, the players involved in a good promotion include: 

• Marketing Managers: Segment targets, identify product categories to promote, create pro-formas, and set timing of campaigns. 
• Marketing Analysts: Pull together the data needed for the promotion, including customer and product records.  
• Sales (including branch managers): Review plans, set sales forecasts, provide input on product selection, pricing, etc.  
• Product Managers: Specify SKU selection by target segment and engage in vendor negotiations for co-op support and pricing, unless there is a separate pricing department, which is often the case.  
• Inventory Management: Review stocking requirements and how to allocate inventory across various locations. 
• Finance: Align financial planning, oversight and measurement to ensure the campaign achieves its goals. 
• Product Data: Maintain the product information management (PIM) database so that the promotion has accurate and robust data, including specifications and images.  
• Designers: Use the PIM and design software to create the pages containing the products for the promotion. 
• Copywriters: Write marketing copy – all the words except for the product descriptions, which should come from the PIM. 

That’s quite a team of people who put in many hours of effort to plan one campaign. Of course, despite all this talent and hard work, some promotions fail to achieve their goals.  

Promotions with AI in 2028 

Now, let’s imagine how things might work four years from now when major distributors have robust generative intelligence models overlaying their systems, with complete access to all data – customers, products and historical transactions. 

The CMO writes a carefully engineered prompt: 

We want to run a fall product promotion with a goal of generating $20M in sales that are incremental over baseline, achieving a 24% gross margin. The promotion will start on October 1 and end November 30.  

Internal Communication 

• After you have evaluated this complete prompt, write an internal memo explaining the promotion to all affected employees. 
• Write a second email giving a general overview of the program to our suppliers.  
• Do not refer to the financial goals in either memo.  
• Prepare three versions of each email and schedule them to be sent to both groups in the 30 days prior to launch. 

For every individual contact at every customer, do the following: 

• Analyze their historical purchases 
• If a contact does not have enough purchase history to recommend a custom set of SKUs, choose a set of 200 SKUs to be used as a default and apply it as necessary. 
• Pick fall promo SKUs that are likely to be purchased by the buyer based on historical behavior and price elasticity of demand.  
  • The average number of SKUs per customer should not materially exceed 200 on average with the count for outliers not exceeding 300.  
• Set pricing for each SKU taking into account profit goals and price elasticity of demand. It’s okay to change pricing between accounts but all contacts within the same account must have the same pricing.  
• Once the SKUs have been selected for all customers, evaluate the probability of the promotion making its overall goal of $20M in sales and 24% in gross margin. If the probability falls below 80%, stop and report what financial goals we are likely to meet at that probability threshold. If it’s higher than 80%, proceed to the next step. 

When SKUs have been selected for each customer, do the following: 

• Construct two general promotional designs – one for use on our website and one we can print on-demand.  
• Create marketing copy to explain the overall promotion 
• Create Terms and Conditions that comply with our policies and that are based on our historical Ts & Cs. 
• Create a custom email for each contact, utilizing the SKUs selected for them.  
• Check with suppliers to ensure that sufficient quantities of inventory are available to support the campaign. Where necessary, replace out-of-stock SKUs with available SKUs. Do not miss the deadline for launching the promotion.  
• Check for related items: Make sure complementary or essential SKUs are properly merchandised together.  
• Populate each email with the SKUs selected for that contact and apply pricing. 
• Complete the layout of each email in a way that promotes the most-likely-to-sell SKUs on the first page. 
• Based on all historical marketing campaigns, re-run the analysis to predict the sales and margin of this campaign by calculating the projected revenue and margin of each individual email and aggregating the total. If the confidence is less than 80%, stop and report what financial goals we are likely to meet at the probability threshold. If it’s higher than 80%, proceed to the next step.  

Order Inventory 

• Once the previous steps are completed to the 80% confidence interval, cut POs to suppliers to bring in stock. Monitor shipments and notify me of any shortfalls that will reduce the impact of the program by more than 2% in sales or profits. 
• Remove out of stock items from the cover of each contact’s custom flyer and replace with SKUs that will be in stock. 
• Determine stocking levels for all our locations and distribute inventory appropriately.  

Once the campaign is ready to launch, do the following: 

• Create two additional creative themes, switching up the headlines and design and revise the marketing copy. Apply this to 10% of the recipients, chosen at random.  
• Send 25 examples of each creative execution to me for review and audit.  

Once I give you permission to proceed, do the following: 

• Send each of the creative executions to 1,000 recipients chosen at random but ensure they each represent the full audience.  
• Follow email best practices and guidelines to optimize deliverability and clickthrough.  
• Evaluate the performance of the three creative executions and re-run the financial pro-forma and report this analysis to me.  

When I instruct you, use the creative execution that performed the best to launch the campaign to the rest of our customers.  

• Spread out the emails over a 72-hour period so we can monitor for problems.  
• Once again, comply with email best practices.  
• Update me daily regarding the performance of the campaign and notify me immediately in case: 
  • The campaign is performing at 95% or less of its goals. 
  • The campaign is performing at 110% or more vs. its goals.  
  • Open and clickthrough rates fall more than 150 basis points below our historical norms. 
  • A greater than average % of these emails are being marked as spam. 

What Could We Expect from This Campaign? 

• First, you’d expect superior results simply because this campaign is so targeted – every customer receives a custom promotion with SKUs and pricing selected specifically for them.  
• Second, you’d expect the AI model to produce better results over time – after all, AI is the world’s first technology that learns and improves on its own. This should show up in the results. 
• Third, the distributor could put together promotions in much faster cycle times. The AI system could do all the work above much faster vs. their human counterparts.  
• Fourth, distributors wouldn’t need as many people to operate. Every individual involved in this program has other responsibilities, of course; no one who works on promotions only does promotions. But AI will eventually be capable of displacing much of the non-promotional work, too. This has serious implications on the employment and career of a large group of people, but it’s a huge win in terms of lowering costs and improving profitability.  

How Likely is This Scenario? 

It’s not “if,” it’s “when,” at least in my opinion. The tools are under development – GPTs are already available to overlay your technology stack. These systems have much to learn about your products, customers, historical transactions, etc., and they currently aren’t ready to do the work necessary to create “segments of one” custom promotions.  

But adjacent systems already exist for every task required to produce the promotion I described above (PIM, Marketing Automation, etc.) and these are quickly becoming AI-enabled. Eventually, GPTs that can drive and coordinate the work of multiple systems based on your data will come online.  

What’s the Current State of AI in Distribution? 

We have two upcoming programs to help you understand AI in distribution today: 

2024 State of Distributor Technology: Building Your Tech Stack for Success. This virtual program, hosted by our own technology expert, Jonathan Bein, Ph.D., covers how AI is affecting applications across the distributor technology spectrum. Sign up to access the recording; it’s free.  

Applied AI for Distributors, June 4-6, 2024, at the Chicago Marriott O’Hare: Join other thought leaders at the biggest and most exciting live event dedicated to AI in the distribution industry. Leading speakers and technology companies, along with some of the most technology-enlightened executives in distribution, will come together for unprecedented learning and networking. Learn more.  

Ian Heller

Website

Ian Heller is the Founder and Chief Strategist for Distribution Strategy Group. He has more than 30 years of experience executing marketing and e-business strategy in the wholesale distribution industry, starting as a truck unloader at a Grainger branch while in college. He’s since held executive roles at GE Capital, Corporate Express, Newark Electronics and HD Supply. Ian has written and spoken extensively on the impact of digital disruption on distributors, and would love to start that conversation with you, your team or group. Reach out today at iheller@distributionstrategy.com.