The amazing thing when you look closely at the food distribution industry, and the food sector as a whole, is how incredibly difficult it is to keep everything running. What is even more amazing is that despite these difficulties, we are still able to find the food we want at supermarkets or order the dishes we like at a restaurant.
This week we wanted to share our perspective on the core challenges that hold food distribution back, some of the root causes behind them, and what it looks like for distributors when those barriers start to come down. Underpinning all this is the fact that technology and process changes have created new possibilities that were previously difficult to imagine, opening the possibility that the incredibly difficult work of food distribution becomes slightly less difficult.
The overarching problem that makes food distribution so difficult is margin. Every challenge in the industry starts with the reality that food distribution is a very low margin business and every problem can be summarized as things that put pressure on margins.
Rising labor costs, freight, energy, and packaging all eat into those margins before a distributor has done distributing. The problem then becomes that thin margins leave no buffer for disruption, which is a reality that impacts almost every decision an operator makes.
At Burnt, we categorize the industry challenges into three main buckets, which are: structural challenges, people challenges, and process challenges. Even though many of these challenges are caused by factors external to distributors, we believe all of them have practical solutions that companies can implement today.
Highest on the list of structural issues are general market conditions. For US distributors, the last two years have been defined by tariff volatility and commodity price swings that make it difficult to maintain accurate quarterly forecasts, let alone any kind of long term planning.
The second major structural trend is consolidation, and the clearest signal of this is Sysco's pending 29.1 billion dollar acquisition of Restaurant Depot (we spoke more about this in our newsletter here). Consolidation at this level affects market dynamics where it becomes increasingly difficult for distributors to compete on price alone, though as we mention in our previous newsletter, there are other ways to compete. The fact that the Independent Restaurant Coalition has asked the FTC to block the deal is a signal of how impactful this is expected to be for the industry.
The third structural challenge is the evolving regulatory environment. Regulations across different markets require new traceability records, sometimes with physical checks on products, including imports. Compliance cost is a major burden, and for mid-market distributors it is one that can have an oversized impact.
At first glance, these look like issues beyond the control of any individual distributor. That is partly true. Tariffs will keep moving, consolidation will keep happening, and regulations will keep arriving. But how a company responds to these shifts can be very much within its control. And in 2026, the tools available to respond are very different from what they were even five years ago.
The practical response has two parts. The first is de-risking through better data, across supply chain planning and pricing. In the past, supply resilience was largely relationship-driven. A good buyer differentiated through contacts and suppliers that competitors didn’t know about. That advantage is getting rarer as sourcing data becomes more accessible. Today, resilience increasingly comes from being able to model supply alternatives across products, and geographies quickly, and from having specific pricing playbooks ready for specific scenarios.
This shifts the work from mapping supply chains to something more like data science, running forecasts, creating scenarios and running these scenarios proactively when data signals indicate a shift is coming. The real problem for many distributors is that this work requires good data. The better the data, the better the analysis, including being able to have a full history of actual costs per SKUs, per route, per customer and how external conditions have impacted each of them. Bad data means that distributors lose one of their most powerful weapons, which is the years of history across operations, purchasing, sales and costs that can be used for better adaptability to future scenarios.
Part of how to implement this is process driven (making sure that ERP data is clean, well maintained and well structured, having efficient workflows for different teams). But AI tools are also adding a very powerful toolset to solve structural challenges. It does this in two main ways. First AI tools can process data in any form (structured and unstructured) meaning it can help close the gap on bad data from ERPs and other systems, without having to get all the data in a perfectly structured format in the first place. AI is also very good at identifying patterns and trend analysis. It can flag very quickly when it notices specific signals and flag these to teams to take action. In both cases, the right technology approach can meaningfully shift the type of work within a distributor from reactive to proactive, which ultimately the only real tool companies have to navigate externally changing conditions.
The second major challenge is labor. This breaks down into two connected problems.
The first is the shortage itself. In the day-to-day of a distributor, the shortage of labor is an issue that affects all areas of operations, shortages of drivers, high warehouse turnaround, fewer new hires for operational roles. This materialises itself into outcomes like having more routing disruptions and late deliveries if drivers get sick, forcing dispatchers to reroute in real time to cover gaps. Or overtime costs grow as existing staff cover gaps in the warehouse and decreasing customer service quality as leaders balance key team members retiring and the challenges of quickly training new employees from a decreasing labor pool. All of which drives burnout, more turnover and more difficulties in maintaining business standards and results.
The second problem is the knowledge loss that comes with that turnover, particularly in roles outside of driving. In food distribution more than most industries, the best people carry decades of context in their heads. The buyer who reads a small price move on pork bellies in Iowa and knows this signals to lock bacon contracts before the next auction cycle. The CS rep who knows what each chef orders and can follow up if they notice a change in order patterns. The warehouse lead who knows which routes work when the weather turns. This knowledge rarely lives in any system. When the people who have this knowledge leave, the knowledge walks out the door with them, and the cost lands on higher training costs, more error rates, and worse customer experience at once.
Some of this might eventually be addressed by things that are beyond our direct expertise like warehouse robotics, autonomous vehicles, and other capital-intensive automation, but in the context of food distribution, a lot of this is still in the realm of science fiction.
The practical response available to distributors today is different, and it comes in two parts.
The first is doing more with less. The current news cycle talks about AI replacing jobs. This is not matching the reality on the ground, where what is actually happening is AI unlocking more productivity per worker, especially in industries dominated by manual labor. More on this in the next section.
The second is capturing institutional knowledge in systems before it walks out the door. This does not mean asking senior employees to document everything, which rarely works. It means building systems that learn from actual work. These are systems to learn customer preferences surfaced from order history, exception-handling patterns captured from CS interactions, supplier reliability scored from real data. The capabilities to do this exist, made possible again by AI systems that learn from user interactions and build a memory of how people work and use their judgement to take action. The 30-year buyer's judgment starts to live alongside the tools in a way that can be shared with the next generation rather than leaving with them.
What does the outcome look like? Routes are optimized enough that fewer drivers cover the same territory. New hires become productive in weeks rather than months because the context they need is captured in systems. Institutional knowledge amplifies the whole team rather than being trapped in the heads of a handful of veterans.
The third area is process: how work actually gets done day to day.
This is the category most visible inside a distributor, and the one most directly within a company's control. The core problem is that historically, food distribution has scaled linearly. More orders means more people. There is little to no natural efficiency gain with volume. Every new customer adds cost at roughly the same rate as they add revenue. That is the operating leverage problem, and it is the reason mid-market distributors often struggle as they grow. Volume is supposed to make you more efficient. In food distribution, volume often just means more work.
The clearest example is order processing. Orders still arrive by email, voicemail, fax, texts, and handwritten notes. CS teams re-key every one into the ERP. At scale this is massive. We have worked with a meat distributor where customer service was spending the majority of its time on manual order entry and exception chasing before automation. Every increase in volume required more headcount. Cost per order stayed in dollars when it should have been in cents.
The ripple effects are significant. Customer service stretched thin on data entry has no time for the relationship work that actually drives retention, let alone growth. Exceptions get handled reactively rather than proactively. Upsell conversations do not happen. Customers notice, although they usually do not tell you before they switch.
Demand forecasting is the second major process challenge. Most distributors still rely on spreadsheets and gut feel. This either works just well enough, or more likely, it leads to over-ordering and under-ordering, both of which have their direct costs.
Pricing is a third process gap. Two things make dynamic pricing rare in distribution, and they are easy to conflate. First, most distributors lack the real-time cost data that would let them price dynamically. Second, even when they have the data, many choose not to change prices frequently because they are worried about eroding customer trust. Both concerns are legitimate. But they interact badly: distributors absorb cost increases for too long, then raise prices all at once when the margin pressure becomes unsustainable, which is exactly when customers notice and get upset. Dynamic pricing, done well, is not about changing prices every week. It is about knowing precisely what each customer costs to serve, where the margin is coming from, and timing increases when the case is defensible. Sales reps quoting without margin visibility make this impossible.
The practical response starts with where to focus. The answer for most distributors is the same: automate the highest-volume, most repetitive workflow first, which is usually order entry. AI-powered order capture reads orders from any channel, validates them against the catalogue and business rules, and pushes them into the ERP. Humans only see the exceptions. The same pattern extends to PO matching, invoice reconciliation, and status updates: automate the routine work, keep humans on judgment calls.
From there, the redesign extends to more process driven and organizational work. CS roles refocus on customer outcomes rather than process execution. Demand forecasting uses historical data plus external signals like weather and seasonality rather than instinct alone. Pricing gets dynamic rules built on input cost, margin target, and competitive positioning, with real-time visibility for sales reps at the point of quoting.
What does the outcome look like? The meat distributor we mentioned earlier now has 97 percent of orders processed automatically. Customer service headcount has been redirected significantly toward higher-value work rather than expanded to keep up with volume. Forecast accuracy climbs meaningfully when historical data is combined with external signals. Waste drops in the same direction. Pricing becomes a strategic tool rather than a guessing game. Margin leakage that was invisible before becomes visible and gets fixed. And in the end, all of this really means one thing: better margins.
When structural, people, and process challenges start to be addressed together, the math of the business changes. The exact numbers vary by starting point, but the direction is the same.
More importantly, the business changes character. Volume grows without headcount growing at the same rate, which is the operating leverage food distribution has never had. The 30-year veteran's expertise amplifies the whole team rather than sitting locked in one person's head. Customer relationships deepen because the people handling them actually have time for the work. Compliance is handled in the background rather than requiring a dedicated function. Mid-market distributors stop being squeezed by the Big 3 on one side and regional undercutters on the other, because they compete on intelligence about their customers, their suppliers, and their markets, combined with superior service, rather than just scale. Independence becomes sustainable again, which matters when the alternative is selling to PE or the Big 3. But even more, it becomes an opportunity to build differentiation and gain an advantage by providing better service and better products for improved customer satisfaction and better relationships.
This is becoming less and less theoretical every day. The solutions are emerging, and a growing number of distributors have started using them. In the end the food will always find a way to reach our plates, but the future we see is one where the people getting it there will struggle a little less.