Breaking down data silos: a conversation with John Manlove
The cannabis industry does not have a shortage of data. It has a problem connecting that information and turning it into decisions operators can use.
That distinction has shaped the development of Apex Trading, a cannabis wholesale platform operating across 30 states. Rather than attempting to own every part of the technology stack, Apex has focused on opening its ecosystem and integrating with the tools its clients already use.
“We don’t have a data problem in cannabis where we need more data,” Apex Trading CEO and co-founder John Manlove told Greenway. “What we need is to end the silos, break down the barriers of data, centralize it, standardize it, make it uniform, and make it actionable.”
That approach extends from how Apex develops products to how the company views artificial intelligence, purchasing behavior, accounts receivable, and one of the industry’s most persistent problems: oversupply.
Growth through product development
Apex has expanded without building the large staff commonly associated with national technology companies. Manlove noted that the company has not raised capital in more than four years, relying instead on product-led growth and a relatively lean team.
“We really believe, and continue to believe, that if you build the right product, listen to clients, continue to improve that, and expand your integrated ecosystem, generally, network effects will result in growth,” he explained.
That philosophy also requires a strong service component. Manlove said personal relationships remain foundational within cannabis, particularly when technology providers are serving operators under significant financial pressure.
“A really good product that people speak highly of and refer will lead to growth,” he continued. “And then, ultimately, a team that’s very passionate, that really shows that they care, that they listen, answer the phone, and show up to support clients, builds kinship and relationships, which we know are the bedrock of this industry.”
For smaller cannabis businesses, that attention can provide a way to compete against more heavily capitalized companies. A craft producer may not have access to every retail chain, Manlove observed, but it can distinguish itself through product quality, pricing, customer support, brand identity, and data showing why its products belong on shelves.
“I don’t need to be in 100 stores, but maybe if I’m in 20 stores, I want to be the best vendor in those 20 stores,” he remarked.
Being the best vendor involves more than delivering products. It means helping retailers understand demand, carry appropriate quantities, prevent products from expiring, and respond when inventory does not move as expected.
“It isn’t only delivering consistent quality, well-priced product that moves,” Manlove elaborated. “It’s also working with that retailer to understand: How do we get the right consumers in the door? How do we collaborate to make this a really successful partnership? When product doesn’t move, how do we help get it moved? How do we make sure that you’re ordering the right quantity at the right times based on true demand?”
Meeting buyers where they are
Apex has developed several ways for brands and retailers to transact, including custom e-commerce storefronts, embedded website menus, a wholesale marketplace, and live Google Sheets menus powered by its application programming interface.
The options reflect a basic reality: Buyers do not all want to order the same way, and sellers do not all want to sell the same way.
Across Apex’s 30-state footprint, Manlove said buyers create approximately half of all orders. The percentage varies dramatically by market, ranging from less than 10% in some states to between 80% and 100% in others.
When Apex examined why buyers were not creating more orders through the platform, the answer was not necessarily dissatisfaction. Some purchasing managers overseeing multiple stores simply found it faster to work from a spreadsheet.
Manlove recalled that buyers identified several problems with spreadsheets: They do not update live, they may not reflect custom pricing, and managing them can be arduous. Rather than attempting to force those buyers into a particular workflow, Apex developed a way to retain the familiarity of Google Sheets while addressing those shortcomings.
“Rather than us saying, ‘Let’s force buyers and brands into our marketplace and into our e-commerce, where we monetize the data and their behavior,’ we said, ‘Screw that,’” Manlove recounted.
“Let’s open it up where Apex can be the backbone, but we don’t have to be the arms and legs. If that’s the arms and legs you want, we can be. But if you don’t want to, let’s not be selfish and self-serving.”
Manlove reported that nearly 1,000 orders per week are now created through Apex’s Google Sheets import or its API.
The company is also integrating with other purchasing and inventory tools, allowing clients to work from the system that best fits their operation while Apex supports the transaction behind the scenes.
“We would love for buyers to be in our app, but we’re not going to force it,” he emphasized. “We’re going to meet the market where it is. We’re going to open it up, understand these patterns of behavior, and, through partnerships, we think we can start to improve.”
Better ordering can improve demand signals
Making wholesale ordering more efficient is not only a matter of convenience. Manlove believes it can help address the inaccurate demand signals that contribute to oversupply.
A retailer holding 60 days of inventory may appear to be demonstrating stronger demand than actually exists. If purchasing tools help that retailer carry between 14 and 30 days of inventory and place smaller orders more frequently, products may sell through faster and require fewer discounts.
“If a buyer is using these optimization tools and ordering based on 14 to 30 days of inventory on hand rather than holding 60, they’re selling through their product like traditional CPG,” Manlove explained. “They’re not holding inventory, they’re not discounting expiring product, and they’re not doing credit buybacks.”
More importantly, cultivators and manufacturers receive a more reliable indication of what the market is consuming.
“They’re giving the brand a true demand indicator so that the cultivator with an 11- or 14-week cycle is actually producing based on what that true demand is, rather than overorders and overpurchases,” he added.
Cannabis cultivators may work on production cycles extending 11 to 14 weeks. When retailers overorder, producers can mistake those purchases for sustainable demand and increase production accordingly. By the time the product reaches the market, retailers may still be carrying inventory from previous orders.
That cycle can result in discounting, declining wholesale prices, compressed margins, and unpaid invoices.
“We’re not going to solve overproduction, but together we can start to improve the indicators and the behavior,” Manlove acknowledged. “Over time, cultivators, who are furthest from the demand signal, can start to produce based on the true demand in their state.”
AI creates opportunity, not certainty
Artificial intelligence is playing a growing role in Apex’s internal operations and the systems used by its clients.
Manlove said early attempts to use AI in the company’s development process produced unreliable results. As the technology improved and Apex established stronger rules and safeguards, it became considerably more useful.
According to Manlove, the company has experienced a 1,400% increase in development output after incorporating AI into engineering, code review, and quality assurance.
“We’re at a 1,400% output increase since we’ve been able to refine and build our model on the development side,” Manlove reported. “We’re shipping more features and tools that are more sophisticated, more enhanced, and have better code than we ever have.”
Apex is also using the technology to identify patterns in customer behavior, measure feature adoption, recognize signals that a client may require assistance, and better understand what different types of operators need from the platform.
“We can create matching personas. We can track feature adoption,” Manlove detailed. “We understand what features drive great clients and what features drive churn. We can understand what a small cultivator in Maine loves and uses versus one in Oregon because they’re not the same.”
The company is beginning to apply similar tools to sales, onboarding, and customer success. By identifying which features are most useful to comparable businesses, Apex can focus training and onboarding on the tools most likely to provide immediate value.
For operators, AI can help combine information from wholesale platforms, point-of-sale systems, enterprise resource planning tools, scales, accounting software, and marketing platforms. But Manlove cautioned that connecting those systems without proper governance can generate unreliable or damaging results.
“We’re all running quickly with AI, maybe without slowing down first and making sure the whole house is in order,” he stated. “Sometimes it’s really important to slow down, create the right governance, create the right protocols, create the right rules, and make sure the data coming in is clean because it’s junk in, junk out. Then you can start to get the outputs you need from it.”
A purchasing tool, for example, may identify an item as a fast seller without recognizing that its velocity increased because the product was nearing expiration and heavily discounted. If the system recommends another large order based on that incomplete conclusion, the retailer could compound the original inventory problem.
“Maybe it’s not accounting for expiration dates,” Manlove offered. “Maybe it’s not accounting for the fact that the fast mover last month was expired, we discounted it, and that’s why the velocity was high. We actually don’t want to reorder that much. It was one factor that it didn’t account for.”
He cautioned operators against trusting an automated purchasing tool without retaining human oversight.
“That’s my concern, that you start to implicitly trust these things without having that human loop, the one who is questioning every single thing,” Manlove warned.
He also advised operators against assuming that an AI tool can quickly reproduce a platform that specialists have spent years building, testing, and maintaining.
A business may be able to create a functional dashboard or purchasing tool, but it must still determine who will maintain it, validate its outputs, correct errors, update integrations, and accept responsibility when something goes wrong.
“We have a saying here in our company, which is, ‘Focus is as much about what you aren’t doing as what you are,’” Manlove shared. “If you’re focusing now on building technology that you’ve never built before, are you shifting away from the focus on your craft and the other elements of your business? Are you distracted?”
Hidden cost
That same calculation applies to systems operators may perceive as free.
Google Sheets carries little direct expense, but Manlove said businesses often overlook the labor required to manually enter orders, reconcile changes, communicate between buyers and sellers, and update inventory.
He offered the example of two sales representatives earning approximately $35 per hour after commissions and spending eight hours each week manually entering orders. That labor alone could cost the company roughly $2,500 per month.
“Sheets aren’t free,” Manlove stressed. “They are not free.”
More significantly, the sales representatives are spending that time as order takers instead of generating revenue.
“What you’re paying your reps to do is be order takers and importers rather than revenue generators,” he explained. “I don’t hire sales reps to maintain my existing accounts. I have account managers for that. My sales rep is opening new doors and closing new deals. That is their job.”
If buyer-created orders reduce the amount of manual entry by half, the company can redirect that time toward preventing account churn, expanding existing accounts, and reaching new retailers.
“Not only are we going to save that labor, what is that rep going to do during those eight hours?” Manlove posed. “They’re going to generate new revenue. They’re going to open new doors, prevent churn, and increase SKU count at existing doors.”
A spreadsheet that appears free may ultimately cost tens of thousands of dollars annually when labor across sales, inventory, accounting, and reconciliation is considered.
“There are so many impacts when you start to really break it down,” Manlove reflected. “A lot of companies say, ‘This is just what I’m on, and it works.’ But do you understand the true cost of that thing?”
The AR problem
Operational inefficiency becomes even more dangerous when combined with the cannabis industry’s growing accounts receivable problem.
Citing estimates from Whitney Economics and FundCanna, Manlove said delinquent cannabis accounts receivable increased from approximately $3.8 billion in 2023 to $4.2 billion in 2024, with a more recent FundCanna estimate approaching $5 billion.
“When you look at that, that’s almost 20% of the entire industry’s overall revenue,” Manlove calculated. “When you look at the valuation of an industry, it’s in AR. That is crazy. It’s significant.”
He said cultivators and manufacturers are especially vulnerable because they effectively finance retailers through extended payment terms, often without standard credit protections.
“The growers and manufacturers are the biggest financiers in the cannabis supply chain,” Manlove asserted. “They’re just doing it with no underwriting, no contracts, no interest, or anything else.”
Small and minority-owned operators may absorb the greatest damage because they generally have less capital available to withstand invoices that remain unpaid for 90 days or longer.
“When we think about the AR issue, who is the most impacted? The growers,” Manlove explained. “Of those, small and minority operators can’t survive it. The people who have taken the most risks in the interest of the industry are the most impacted by the growing AR problem.”
Competitive pressure can make the situation worse. A brand may continue selling to a retailer that regularly misses its payment terms because refusing the next order could allow a competitor to take its place.
“A brand knows, ‘Maybe I allow a buyer, as long as they pay me. I know I gave them net 30, but as long as they pay me in 90, I’ll keep selling to them,’” Manlove illustrated. “Because if I stop selling to them, somebody else is going to come in, and I’ve just lost that revenue. They value the revenue more sometimes than the payment, and that’s a big issue.”
Apex is working to incorporate credit information and payment risk into the wholesale process so brands can evaluate a retailer before extending terms. The company has also integrated FundCanna’s ReadyPaid service, which allows an approved retailer to receive net-30 terms while the brand is paid upon delivery.
“The retailer gets net 30 with no interest, and the brand gets paid on delivery, where they don’t have to worry about collections,” Manlove outlined.
Those tools cannot solve the problem alone, but greater transparency can help operators make informed decisions about whether to offer net terms, require payment on delivery, or decline an order.
“We’re trying to bring in solutions and partners to solve it,” Manlove clarified. “We know we can’t solve it ourselves, but we know today that AR is the extinction event for so many operators.”
He has seen companies enter Apex with hundreds of thousands of dollars in outstanding receivables, much of it more than 90 days overdue.
“We’ve brought clients in who say, ‘I’ve got half a million in AR.’ How much of it is past 90 days? ‘$300,000.’ You’ll never get it,” Manlove recounted. “That is a death wish for a lot of operators.”
Large orders and multi-store chains can be attractive, but volume has limited value if the retailer does not pay. In some cases, he said, brands may be better served by smaller independent retailers that order less but pay on time and actively support the products they carry.
“The data shows that the biggest retail chains, the biggest operators, are the worst payers,” Manlove stated. “The people with the most money and the most capital are the ones who pay less frequently and don’t meet the terms as well as others.”
A brand may view placement in a 20-store chain as a major opportunity, but Manlove cautioned that projected sales should not overshadow payment history.
“Maybe I’m better off being in these small stores that pay me on time and support my brand,” he reasoned. “It’s less volume and smaller baskets, but, at the end of the day, we’re not delivering a service we’re not being paid for.”
Data to correct oversupply
For Manlove, the cannabis industry’s largest opportunity in data is correcting oversupply.
Excess inventory affects nearly every part of the market. It lowers wholesale prices, forces retailers and brands to discount products, distorts production planning, reduces profitability, and contributes to unpaid invoices.
“For me, when I look at it through our lens, it really comes back to fixing oversupply,” Manlove answered. “That today, in wholesale, is where my goal is. It’s not only AR, but really the oversupply issue because that is the reverberation and the trickle-down of everything.”
The solution is not simply producing more data. It is connecting retail demand, inventory levels, wholesale ordering, production planning, and payment information so businesses can act before problems spread through the supply chain.
Manlove believes retailers should carry less inventory, order more frequently, replace slow-moving products sooner, and give producers more accurate signals about actual demand. In turn, producers can adjust output rather than continuing to chase sales with excess supply and lower prices.
“How do we get retailers to carry 14 days or 21 days, order more frequently, order more often with lower baskets, but sell through it?” Manlove asked. “They’re keeping products in, buying the right products, and replacing the slow movers.”
“When did scarcity become really bad in the industry as a producer?” he continued. “Scarcity is a good thing. You want that. You want to be able to say, ‘Sorry, we’re sold out. Next time I send you the menu, place the order right away.’”
After years of building their respective platforms, Manlove believes Apex and its technology partners now have the infrastructure needed to begin connecting those pieces at scale.
“We’ve all been building these planes as we’ve flown them, and now we’re at altitude, cruising, and everything is really stable,” he concluded. “A lot of us that have been around for seven or eight years are now at the phase where we can achieve this. We can actually do this, but we’re going to do it together.”
