A ChatGPT Experiment: 13 key takeaways for B2B SaaS content marketers

At the end of the day, the companies that will experience the greatest success with AI will do so because they have the right attitude. They understand that AI shouldn't be used to replace humans. Rather it should be used to augment them. To elevate their work. And ChatGPT is no exception. So how can you use it to augment content marketing? And what are some pitfalls to watch out for? Let’s take a look.

A ChatGPT Experiment

Below, I’ve documented my first ChatGPT journey and shared some observations along the way. Note that it's not essential to read the copy in italics which is the output of ChatGPT (unless you're in retail ecommerce and distribution, you probably won't find it very interesting), but I've included it for reference.

Experiment Goal

See how well ChatGPT can write an article on how retailers can optimize split shipments.

Attempt 1

Input: Describe some ways a retailer can reduce split shipments

Output:?

Split shipments occur when a retailer sends an order in multiple shipments rather than all at once. This can happen due to a variety of reasons such as out of stock items, warehouse location differences, or shipping limitations. However, split shipments can be costly for retailers as it increases shipping costs and can create a poor customer experience. Here are some ways retailers can reduce split shipments:

Inventory management: Proper inventory management is essential to prevent out-of-stock situations. Retailers can use forecasting tools to estimate demand and ensure that they have sufficient stock of popular items.

Centralized warehousing: Having a centralized warehouse can help ensure that all items are in one location, which reduces the likelihood of split shipments due to item unavailability.

Stock levels: Retailers can monitor the stock levels of each item and set thresholds for when to reorder. This ensures that inventory levels are always maintained, and popular items don't run out of stock.

Order management system: Retailers can use an order… <ChatGPT stopped at this point>

Observations

• The right input is essential - The input wasn't very good. I had used the word 'reduce' rather than 'optimize'. And I had a very specific idea for what would be returned based on my 'retail order fulfillment' lens, but ChatGPT interpreted my question, and answered it, at a higher level than was expected. The input needed to be more specific. What I wanted to focus on?the role of an order management system in reducing split shipments?was but one bullet point in a longer list.
• Provides good buyer persona perspective - ChatGPT helps you understand the broader context of what you're asking. The output serves as a good reminder that your buyer personas aren't just focused on the problem you solve, or how you solve it.They have a lot of different inputs to consider.

Attempt 2

Input: Describe some ways an order management system can help a retailer optimize split shipments.

Output:?

An order management system (OMS) can help a retailer optimize split shipments in several ways:

Real-time inventory tracking: An OMS can track inventory levels in real-time, helping retailers determine which items are in stock and which are out of stock. This information can be used to optimize split shipments by ensuring that only in-stock items are shipped.

Intelligent order routing: An OMS can intelligently route orders to different warehouses or distribution centers based on factors such as inventory levels, shipping speed, and cost. This can help retailers optimize split shipments by sending items from the closest warehouse to the customer, minimizing shipping time and cost.

Efficient picking and packing: An OMS can optimize split shipments by streamlining the picking and packing process. By grouping items that are going to the same customer or location, the system can reduce the number of times a picker needs to go to the same location to pick items, thereby saving time and reducing the likelihood of errors.

Order prioritization: An OMS can prioritize orders based on factors such as shipping time, shipping cost, and order value. This can help retailers optimize split shipments by ensuring that high-value or time-sensitive orders are shipped first, while lower-value or less time-sensitive orders are grouped together to reduce shipping costs.

Real-time order tracking: An OMS can provide real-time order tracking to both retailers and customers. This can help retailers optimize split shipments by enabling them to track the progress of each order in real-time, and make adjustments as needed to ensure timely and accurate delivery.

Observations

• Output broader than expected - Once again, the output was broader than initially expected. I'd expected something focused on advanced sourcing rules that can be used to optimize split shipments. But ChatGPT looked at all the different ways an order management system can help optimize split shipments.
• Provides competitive positioning insights - Given ChatGPT's output is based on existing web content, it can provide a quick way to understand how competitors and other experts in your field position the value of your type of solution.

Attempt 3

Input: How can sourcing rules be used to optimize split shipments for retailers?

Output:

Sourcing rules can be used to optimize split shipments for retailers by determining which items should be shipped from which locations to minimize transportation costs and delivery times.

Here are some steps that retailers can take to use sourcing rules to optimize split shipments:

Analyze the inventory: Retailers should analyze their inventory to determine which items are in stock and where they are located. This information can be used to create sourcing rules that dictate which items should be shipped from which locations.

Define sourcing rules: Sourcing rules should be created based on various factors such as the product demand, shipping costs, inventory levels, delivery times, and the availability of the products. Retailers should consider the weight and size of the items, as well as the shipping carrier's rates and delivery times, when creating sourcing rules.

Use a shipping software: Retailers can use a shipping software that supports sourcing rules to automatically split shipments based on the predefined rules. These software systems can optimize the shipment process, reduce shipping costs, and improve delivery times.

Monitor performance: Retailers should monitor the performance of their split shipments to ensure that the sourcing rules are effective. By analyzing the data, retailers can refine the sourcing rules and make adjustments as needed.

Overall, sourcing rules can help retailers to optimize split shipments by reducing transportation costs, minimizing delivery times, and improving the customer experience. By analyzing the inventory, creating sourcing rules, using shipping software, and monitoring performance, retailers can improve their supply chain efficiency and increase profitability.

Observations

• A ‘How’ questions didn’t yield specific results - While open ended 'how' questions are great if you're working with human teams, they don't necessarily work well for a content generating AI. This was still more 'high level overview' rather than specifics. I needed to find another way to describe the problem.
• A ‘How’ question gave a good benefits summary - While the 'how' question did result in a nice 'benefits summary' at the end, you need to be laser-focused on a specific use case to get good results.?

Attempt 4

Input: What are some examples of sourcing rules used to optimize ship from store?

Output:

Ship from store is a fulfillment method used by retailers to fulfill online orders from inventory stored in their physical stores, rather than from a centralized warehouse. Sourcing rules are a set of guidelines that retailers use to determine which store or stores will fulfill each online order. Here are some examples of sourcing rules that can be used to optimize ship from store:

Distance-based sourcing: This rule sources orders from the store that is closest to the customer, which can reduce shipping time and cost.

Inventory-based sourcing: This rule sources orders from the store that has the inventory available to fulfill the order. This can help retailers move inventory and reduce stockouts.

Capacity-based sourcing: This rule sources orders from the store that has the available capacity to fulfill the order. This can help retailers balance workload and avoid overburdening any one store.

Sales-based sourcing: This rule sources orders from the store that has the best sales history for the product being ordered. This can help retailers optimize sales and improve customer satisfaction.

Priority-based sourcing: This rule sources orders based on their priority, with higher-priority orders being fulfilled by stores that are more capable of meeting the required service levels.

Cost-based sourcing: This rule sources orders based on the cost of fulfillment, taking into account the shipping costs, handling fees, and other costs associated with fulfilling the order from each store.

Seasonality-based sourcing: This rule sources orders based on the seasonality of the product being ordered, directing orders to stores that have the most appropriate inventory for the current season.

By using these and other sourcing rules, retailers can optimize their ship from store operations to improve fulfillment efficiency, reduce shipping costs, and enhance the customer experience.

Observations

• ‘Use cases’ better than ‘problems’ - Bingo! Using a specific use case (Ship from Store) in the input yielded much better results than a challenge (split shipments). Always include the use case in your input.?
• Still need to drill down further - There isn’t a lot of substance to any of the answers.
• Keep experimenting - If you're not getting the results you anticipated, try different ways of describing it.

Attempt 5

Input: Examples of detailed Capacity-based sourcing rules for ship from store optimization

Output:

Here are some examples of detailed capacity-based sourcing rules for ship-from-store optimization:

Store Capacity Rule: Assign orders to stores based on the available capacity of each store. For example, if a store has a capacity of 50 orders per day, only assign orders to that store until it reaches its maximum capacity for the day.

Distance Rule: Assign orders to stores based on their distance from the customer. For example, if a customer lives closer to Store A than Store B, assign the order to Store A to minimize shipping time and cost.

Product Availability Rule: Assign orders to stores that have the products in stock. For example, if a customer orders a product that is only available at Store A, assign the order to Store A to fulfill it.

Priority Rule: Assign orders based on their priority level. For example, if a customer has paid for express shipping, assign the order to the store that can fulfill it the fastest.

Store Performance Rule: Assign orders based on each store's historical performance. For example, if a store has a high fulfillment rate and low error rate, assign the order to that store to ensure a high-quality customer experience.

Time of Day Rule: Assign orders to stores based on the time of day. For example, if a store is less busy in the morning, assign more orders to that store in the morning to balance the workload across stores.

Weather Rule: Assign orders to stores based on the weather conditions in the area. For example, if a store is in an area with heavy snowfall, assign more orders to that store during warmer months to avoid shipping delays and potential safety hazards.

Capacity Utilization Rule: Assign orders to stores based on their current capacity utilization. For example, if a store is only operating at 70% capacity, assign more orders to that store to optimize resource utilization and minimize costs.

These rules can be customized and combined to create a capacity-based sourcing strategy that best fits a company's needs and goals for ship-from-store optimization.

Observations

• A combination of multiple queries yields the best results - As you ask more specific questions, you get better results. While none of them are earth shattering, they are definitely more relevant based on my initial intent.

Now that we’ve looked at the results, let’s look at the readability of the output and some key takeaways.

ChatGPT Readability Scores

Newspapers typically target a grade 9 reading level—for a reason. It increases the likelihood that someone will actually read their content. The Economist targets undergraduate reading level—which makes sense for its audience. The default grade level of these ChatGPT responses were as follows:

• Attempt 1: Grade 12
• Attempt 2: Grade 15
• Attempt 3: Grade 16
• Attempt 4: Grade 15
• Attempt 5: Grade 11

So, if your target buyer isn’t an economist subscriber, chances are you’ll want something easier and faster for your readers to consume. What happens when you specify a reading level in your ChatGPR input? Let’s take a look:

Attempt 6

Input: Examples of detailed Capacity-based sourcing rules for ship from store optimization written at a grade 9 reading level

Sure, I'd be happy to provide some examples of detailed capacity-based sourcing rules for ship from store optimization that are written in a grade 9 reading level. Here are three examples:

Rule 1: Always prioritize sourcing orders from stores with the highest available capacity. This means that if a store has more space available in its inventory, it should be given priority over other stores that have less available space.

Rule 2: When choosing between stores with similar available capacity, prioritize sourcing orders from the store closest to the customer's location. This means that if two stores have similar available capacity, the one that is geographically closer to the customer should be chosen.

Rule 3: Avoid sourcing orders from stores that are already near their capacity limit. This means that if a store is almost out of space in its inventory, it should not be given priority for sourcing orders, as fulfilling those orders could result in stockouts or other inventory management issues.

These rules help ensure that orders are sourced from the most appropriate stores based on their inventory capacity and proximity to the customer, while also preventing potential inventory management issues.

Observations

• Fail! - Not only was the content not nearly as informative, when I double checked the output in Hemingway Editor it scored it as a grade 14, not grade 9.

In short, if you want to provide substance in your content, you’re better off taking the more advanced copy and editing it down yourself rather than getting ChatGPT to do it for you.?

13 Key Takeaways

Potential Benefits

1. Accelerate output - No one likes to start with a blank canvas, ChatGPT can provide that kickstart for your content that will accelerate development, so you can focus on what makes your content stand out, not writing the mundane.
2. Don’t miss key points - Even if you don’t use ChatGPT to create content, you can use it to check you haven’t missed any key benefits that others vendors and experts are highlighting.
3. Understand broader buyer context - It can remind your teams that buyer personas aren't just focused on the problem you solve, or how you solve it—they have a lot of different inputs to consider. Your solution probably only solves a subset of your buyer personas’ challenges, and it's important to understand the broader context in which they operate.
4. Discover new keyphrases - As you refine your questions, you may discover new SEO keyphrases to consider?keep track of them!
5. Understand category positioning - ChatGPT is based on data from the web, i.e., what other people have written about your topic, so it’s a fast way to understand how other vendors in your category are talking about the problem space

Words of Warning

1. Input needs to be specific - To get the right output, you’ll need to provide a very specific input. Keep an open dialog across your team about what works best.
2. Cut off date is September 2021 - At the time of writing, the cut off date for ChatGPT’s content was September 2021, so if you're in a fast moving industry, the information won't be the latest and greatest.
3. SMB Bias - ChatGPT won't include gated content, like some analyst reports and industry research, so default results may skew towards the messaging from a larger pool of SMB targeted solutions rather than the smaller pool of Enterprise solutions.

Potential long term implications

1. ChatGPT may provide an alternative to traditional search engines - If you're a buyer looking for a solution to a business problem, ChatGPT can provide a summarized view of what everyone is saying about it without having to visit lots of different websites. An interesting shortcut for the buyer, but not great for brand impressions?there’s no content attribution.
2. Everyone will sound the same - If everyone starts to use ChatGPT for content creation, we’ll all start sounding the same. At the end of the day it’s fresh metaphors, analogies and perspectives that make content compelling. As we look to augment our output, it’s important we don’t lose sight of what we need to do in order to cut through the clutter.

Best Practices

1. Create guidelines - For acceptable use of ChatGPT generated content for your organization. You don’t want your teams to just copy and paste from ChatGPT directly into a blog post.
2. Edit for readability - So you don’t make your readers work too hard to consume your content
3. Edit for differentiation - If you don’t add your own ‘secret sauce’ to content you’ll sound the same as everyone else and simply won’t be memorable.

Summary

Marketers have some great new AI based tools available to them today. From image editing tools like Removal.AI that quickly removes the background from your subject (it even does a pretty decent job on curly hair) to content writers like ChatGPT. But it’s important to remember that successful marketing requires being memorable, creating an emotional connection, and having your own brand voice. We’re still a long way from AI being able to ‘replace the humans’?and really that shouldn’t be the goal. But if used the right way, these tools can increase your team’s productivity, and elevate their work. That way they can spend less time on the mundane and more time on innovative creative and messaging that will disrupt markets and fuel your future growth.

P.S. If you're a retailer or brand that needs to optimize split shipments, let's chat and see if Fluent Order Management could be a good fit for you.