production

As Artificial Intelligence (AI) ushers in the era of Industry 5.0, transforming every industry and every size of business, it’s impact will be most keenly felt in the realm of workflow efficiency as businesses adopt AI for productivity. Repetitive tasks, long the domain of human workers, are now being automated by machine learning driving large language learning models (LLMs) and intelligent machines, leading to significant time-savings and quality improvements. From jump-starting marketing and messaging to streamlining customer service to optimizing production lines, AI is revolutionizing how businesses operate. No matter what size of business you have, you  can do more with less and focus on higher-level work by incorporating AI in your technology stack. The result? Cost reductions, increased productivity, and stronger competitive performance.

AI for Marketing

AI can be a powerful tool in your sales and marketing toolbox. So far, much of the discussion about AI … at least in the small  business sector … focuses on AI as a content generator. Given how much has been written on this aspect already, I’m going to focus on other areas of sales and marketing benefits. (edit 2024/4/8: I recently did a short talk on AI for content generation, which you can watch here). 

Lead Generation and Qualification

AI can also transform your lead generation and qualification efforts by transforming vast amounts of data into insights you sales team can quickly put to use. Tools already exist that can crawl blogs, social media posts, and online forums to identify users interested in specific topics. It can analyze demographics and past behavior to further refine lead pools, and it can produce lead scores based on interactions with your website, email, and social media presence. This type of work would take sales and marketing teams an extraordinary amount of time, but AI tools can achieve similar results in days or even hours.

Predictive Analytics for Campaign Development

AI can also help you target marketing  campaigns by moving beyond simple demographics and past behavior. It can analyze massive customer datasets to identify hidden patterns and correlations … connections that only the most skilled and experienced data analysts would otherwise pick up on. AI tools can predict which customers are most likely to churn, which are prime for upselling, and which customers will respond best to specific types of promotions or messages. 

Wouldn’t you love to know which customers were least likely to repurchase before offering a discount? Would you love to know the best time to send a re-engagement communication to a past purchaser? These are things AI can facilitate for you.

Sales Forecasting

Forecasting is always going to be part art, part science, but AI tools can make the science aspect of forecasting much, much (much) more accurate. This is  because AI tools can factor in far more variables than traditional approaches like spreadsheets with human interpreters or even business intelligence (BI) tools like Power BI or Qlik.

AI algorithms can analyze historical sales data, market trends, customer behavior, seasonality, and even external factors like economic indicators, local weather, and local or regional politics. The tools can pinpoint subtle patterns and correlations that untrained (or even skilled) data analysts might miss. Even better? AI tools can also pull in data regarding competitor product launches, industry adoption rates, and your own lead generation data. All this adds up to better prediction of potential sales activities. 

Why do you need this type of forecasting? Because better visibility to future changes in sales will help you eliminate pipeline bottlenecks, get ahead of production opportunities, and rebalance your offerings and inventory to reflect changes in demand. 

Beyond Marketing

But now let’s look beyond sales and marketing, to topics that are being discussed extensively at the Fortune 500 level, but aren’t being given enough attention for small and medium sized businesses … even though the tools for many of the things I’m about to describe are already available and affordable for smaller businesses.

AI for Productivity in Operations and Logistics

AI tools can help you optimize and streamline your inventory, improve demand forecasting (see Sales Forecasting above), and improve your route planning, again, b y analyzing complex data. AI can take the historical sales pattern data you are probably using right now to plan your inventory, provide a more nuanced view of seasonality, add in lead times and external factors like weather and politics to predict what stock is needed and when. All this leads to better inventory and production decision-making.

For route-planning, AI can integrate with your warehouse management system, add in constantly changing real-time factors such as traffic conditions, weather, construction, and delivery time windows, and update routes automatically. 

This is already useful data for large logistics companies like UPS and FedEx. But imagine what it can do for your local delivery team or regional distribution efforts.

Data Analytics

One of the things most small and medium-sized enterprises (SMEs) struggle with is numeracy (edit 2024/4/8: Check out this snippet from my presentation on numeracy in manufacturing). Even though businesses of every size are swimming in vast quantities of data today, very few companies employ … or truly know how to use … sophisticated data analysts.

Even with excellent BI tools, most companies fail to optimize their data insights for better decision-making, because the typical IT, accounting, or even marketing team member doesn’t really know how to structure the data models that will give them better insights.

AI tools can help with this. AI excels at extracting valuable insights from massive datasets. Machine learning algorithms wade through all that data to uncover the patterns, trends, and correlations that help you make better decisions. Even better, AI data tools can convert those insights into effective graphics to help your management team “see” the data similarly and get on the same page.

These tools can also assist with fraud and anomaly detection. By establishing baselines for typical behavior within your data, AI can flag deviations and suspicious activity. 

This same anomaly detection is useful for manufacturers, which can use AI to analyze sensor data from machines, detect subtle anomalies in vibration or temperature patterns, and signal equipment malfunctions before they cause costly breakdowns or isolate quality problems before you throw more labor at them.

Administrative Automation

How much time and money does your company waste in administrative tasks? From duplicative data entry to unnecessary printing to missed opportunities, admin is an area ripe for disruption.

AI accelerates document processing, automates routine tasks, and minimizes the tedium of data entry. Using tools like Optical Character Recognition (OCR), AI-powered tools can extract text from images or PDFs, making them searchable, editable, and indexable. AI can then classify documents based on content, route those documents to the right people or departments, or trigger specific activities.

In email management AI can draft responses to common inquiries and free up time for other, more valuable tasks. It can even interpret the intent of emails to suggest meeting times (and share a meeting link to keep you out of the “what time is good for you” back-and-forth), dropping those meetings right into your calendar for you. 

In data entry you can use AI tools to analyze forms, invoices, or receipts, extract information from them, and directly populate your databases or spreadsheets.

Of course, accessing these AI tools assumes that you have at least begun the work of modernizing your tech stack (need help with that? Book a consult now to find out how StrategyWerx can help). 

Competitiveness in 2024 and beyond will hinge on your ability to automate more of your business processes. We are nowhere near the peak of effectiveness for AI-powered software yet, but the tools you can use to begin this process are already available and affordable. If there’s one thing that SMEs should be thinking about right now, it’s how to automate processes beyond marketing. Implement any one of the ideas in this article, and you’ll increase your competitiveness immediately.

The first time I struck out on my own, in the 1980s, I was completely unprepared for the realities of budgets and financial management.  I earned a fair amount of money for someone so young, but it never failed - I always reached the end of one paycheck before I received the next. I  did not treat myself to expensive things but I had no idea how much things cost, so I was constantly underestimating how much money I needed and running out when I needed it.

It may seem surprising, but small manufacturing business owners often do not know how much it costs to produce their products.  Not knowing how much things cost, they are left with disappointing results at the end of each accounting period and in need of cash to finance upcoming projects. The two biggest costs most companies incur are the costs of production and the costs of labor, so lack of clarity and control in these areas can be devastating.

The key to knowing product costs is to use a sound costing method. There is no magically easy way to manage costing - it is a demanding task that requires intelligence and discipline. But if you manage costing effectively, you will find profits increasing within a year as you refine your production, merchandising, pricing, and marketing strategies based on the insights you gain.

Your goal in costing is to understand both the overall profit of the company and also the individual profitability of items. As a business manager you must do bottom up and top down analysis at all times. In this case, the top down (macro) analysis is the overall corporate profitability, and the bottom up (micro) analysis is the performance of individual items. Imagine for a moment that you are experiencing a 1% decline in your profit margin. How do you begin to understand what is causing it? If you only understand your profit margin at a macro level, you will not have the insight you need to tweak profit margin at the micro level.

The simplest way to set product costs is to average overhead and salaries by adding them together and then dividing to come up with a production-cost-per-minute. You then multiply this cost by the amount of time required to produce a particular item. This is a fairly common approach, particularly for companies new to manufacturing, but it is fraught with risk. This approach will only work if:

1. All the people in your production environment have the same skills and productivity
2. All the people in your production environment are paid the same amount of money (now and in the future)
3. You are not doing any process batching that can be performed by lower-salaried laborers

If all three of the above conditions exist, then presumably each minute of production time for each product is worth the same thing. In that case, a production labor rate that is averaged across all production workers will work for you.

If, on the other hand, you have variability in pay, skill, and type of process, you must approach costing in a more sophisticated manner. The approach I am about to describe is a derivative of Activity Based Costing (ABC). It's important to note that it's a derivative, because if you study ABC you will discover a nearly molecular approach to costing that is time consuming and suffers badly from point-of-diminishing-returns. This derivative approach takes into account variability in skills, pay, and tasks without the expensive granularity.

Let's use a jewelry example to illustrate. Imagine that you are making two rings, both of 18k gold, both requiring setting three stones. But Ring #1 involves very careful finishing to preserve detail in the band, and it requires channel setting. Ring #2 has a simple band and does not require special attention in the finishing step, and the stones are prong-set. Does Ring #1 take longer to make than Ring #2? Obviously. But what isn't so obvious if labor costs are being averaged and divided by the minute is that Ring #1 may require a $70,000/year production worker for 90% of its minutes, and Ring #2 may require a $35,000/year production worker for 100% of its minutes. In this case the averaging of production time over minutes no longer works, because it understates the cost of Ring #1 and overstates the costs of Ring #2.

The way to manage costing in this environment is to benchmark and test. The benchmark is set at the beginning of production for a given item. Not the first time it's made or even the second - those production times will always be longer and are likely set by a much higher paid person proving the production approach for a new design. You will keep careful track of the total (not elapsed) time of production for new items, and up to the point that an item is released for actual production, you can allocate those times to R&D rather than cost of goods. Do you have to do that? Not necessarily. Companies that are doing a lot of new production that is very similar to previous production may opt not to do that just because they have standardized and minimized the amount of time spent proving a new item. But companies that spend a lot of time in the pre-release-for-production phase may want to consider this.

Once the item is released for production, a cost estimate (based on pre-production activities) is set in the system for the first run. You will likely run slow on the first production run, but you need a cost in the system. Then, on the second run of the new item, you carefully benchmark the time spent  - the time spent in production and the cost of labor for the various steps of the process. That becomes the standard cost for the item. After this, you don't measure each time you run the item (this is where the departure from ABC comes in - ABC would require careful measurement every time. But the time spent doing this costs more than its ultimate value in knowledge). Rather, you set each item to be re-benchmarked on a 2X or 1X annual basis. When you run the re-benchmark, you will typically find some improvement in the production time, but sometimes you will discover drift in process, shift upward in salaries, or shifting of labor types, and these changes will either be reflected in the new cost of the item or will give you the opportunity to correct undesirable drifts and shifts. The only time you should have to analyze items off their typical 1X - 2X/year schedule is if something major occurs, such as when an urban manufacturing  company's entire production group fled to a competitor and he had to start from scratch with new employees.

The downside of this approach is that it requires careful management of costs. When you are benchmarking a new item, you need to consider the averaged salaries of the type of potential labor at each step, you must remember to schedule each item for periodic review, and you must remember to do the scheduled reviews. The other downside is that it is not perfect, but neither is averaging all salaries across minutes. The only near-perfect approach is ABC, and like I said, that involves so much extra work that the knowledge gained isn't important enough to justify the cost of gaining it.

The upside of this approach is that it gives you excellent visibility to the actual productivity of your line. When you are trying to solve a problem, you can drill down into items and groups of items and see where your profitability is slipping. You may realize that you want to raise the price on just one category or subcategory of products, or that you can modify the design of an item or small group of items to reduce the cost, or that you can afford to hire a few more $35k/year employees and shift tasks into batches, or (the list goes on). Using this approach, small manufacturers gain insight into very interesting things. Some of the things my clients have been able to see using this approach include:

1. One client found that customers had a distinct preference for the more complex items in his line. The more complex items weren't always obvious from the price point, but they were obvious in the costing. So he selectively raised prices on the complex items to increase margin, because the lower margin of those items - plus their comparatively higher volume - was having a disproportionate negative effect on profitability.
2. One client needed to reduce her price points but had to be very careful how she did it because she had just started seeing the profitability she needed and didn't want to reverse it. By analyzing costs, she saw that she had the opportunity to put her lowest price production workers on a whole category of items that enjoyed good sales. She reduced those product costs and selling prices proportionately - thereby maintaining the overall margin (though not the dollars, of course), and was able to promote this new lower-priced line to counter the competitive pressure she was feeling.

These are just two examples, but they illustrate the type of knowledge this approach can bring. I rarely see conditions (no variability) that would make averaging all production salaries the most prudent approach to costing. The benchmark and test approach is the one I prefer, because even though it requires more work to accomplish, it is one of the most important functions of business. The resulting knowledge will help you manage not only the pricing of your line, but also generate better design and development ideas for future products and the evolution of the company.

(c) 2010. Andrea M. Hill

This is a situation we encounter too often in our work at the WeRx Brands. I've turned it into a fictionalized account to protect the perpetrators...

An employee (let's call her Lisa) gets hired by a reputable company following her completion of a degree in computer sciences. She's thrilled, and can't wait to start applying her education to real-world projects. Her new boss (let's call him Bill) has been with the company for 19 years. He's the Operations Manager, a truly nice person, and excited to welcome Lisa to the world of work.

In her first few weeks, Lisa notices that Bill does everything manually. Inventory? Updated in Excel by hand. Reports? Compiled from printed logs. When Lisa mentions a tool that could automate half of this, Bill nods and says, "I've heard of that. Just haven't had time to look into it." A few more times Lisa mentions other processes that could be automated with simple, inexpensive solutions. Each time Bill says, "That's great stuff. Maybe down the road."

Lisa finds herself getting really good at... Excel. She can VLOOKUP with the best of them. But her skills, the ones she's worked so hard to develop, start to atrophy.

She applies for a different job at another company, and gets it. When she gives her notice, Bill is really surprised. He says, "I thought you were happy! We were grooming you for leadership." And Lisa responds, "I just didn't feel like I could grow here."

We see this same story play out across many roles:

• The IT/Sys Admin stuck maintaining legacy systems, or helping an outdated ERP system limp along, and not updating or adopting new technology.
• The marketing generalist stuck with traditional campaign methods and management, and working in a first generation CRM without access to automation, personalization, testing technology, or analytics tools.
• The accounting admin stuck in a bookkeeping mindset, faced with slow closes, no use of forecasting or financial modeling tools, and no automation.
• The sales person stuck with rudimentary CRM (or no CRM at all), not exposed to modern selling methods like data-driven selling and account-based management, and not able to benefit from extensive sales automations and support.
• The HR person stuck in transactional HR, without focus on strategic talent development or employee experience.
  The Customer Service rep stuck in basic problem-solving, rudimentary ticketing, not viewed as a revenue driver, not using chat, knowledge bases, or leading CRM tools.
• Manufacturing supervisors stuck in manual scheduling with low tech-adoption, no automation, no use of lean practices.

I could go on, but you get the picture.

When it comes to people development, we set the pace. When leaders stop learning, the whole team slows down. Technology changes how work is done, and staying current isn't just good for business... it's a signal to employees that their growth matters.

For those of us in leadership roles, our learning curve doesn't end. We're not just responsible for delivering results (though that's reason enough). We're also responsible for creating an environment where people can grow. Falling behind doesn't just limit ourselves — it limits everyone who's counting on us to lead the way. Want to retain good people? Keep learning, and stay worth following.