/

/

Expert Interview: Scott Scharf

Expert Interview: Scott Scharf

Expert Interview: Scott Scharf

Transcript

Chris:
Hello, everybody. I'm Chris Brunelli with Surpass Commerce, and I'm here with Scott Scharf.

Scott:
Hi, everybody.

Chris:
So, Scott, for those who don't know you, would you care to give us a little bit of an intro, a little bit of your background?

Scott:
Yeah. So I'm a very longtime IT geek, but in 2011, my wife, who's an amazing entrepreneur, a CPA, and an amazing accountant, and I started Catching Clouds, the first e-commerce accounting firm. We were right on Xero, right into the cloud. Our first client was an Amazon seller. No other accountants we had ever seen focused on e-commerce, product-based businesses back in those wild, wild west days of 2011. We went all in to become experts on e-commerce founders, the e-commerce tech stack, e-commerce businesses, and of course all the cloud accounting, bill payment, payroll, and all those other things. I was part of advisory councils, and I was speaking at accounting conferences and e-commerce conferences about e-commerce accounting and how to optimize your business, because once you get beyond a hobby and you're running a real business, you have to have detailed, accurate financials.

The whole time we were running Catching Clouds and doing the bookkeeping, accounting, and controller work for sellers, I was providing the same thing I do now, which is helping e-commerce founders be better entrepreneurs and make better decisions. Looking at business frameworks like the Entrepreneurial Operating System, summarizing the many books that Chris and I have both read, pulling those together so you don't have to read them yourself, and showing you how to apply them to your business. Then helping with decisions. And of course my real focus is helping people with organizational and technology change. Most people aren't a tech geek like I am, and don't deeply understand all the different technology, from Google and Microsoft to security tools, accounting tools, inventory management, EDI tools, sales tax tools, all of them. Like, what's an OMS versus an IMS versus a WMS? And how do those relate and integrate with EDI and a million other three-letter acronyms I could throw out there?

But really, I just love entrepreneurship. My key niche is working with e-commerce entrepreneurs who have their own BHAG, their own big hairy ass goal, and their own goals personally and professionally, and who have decided to take on the risk with their kids' college funds and their home equity lines to start these businesses. How can I help them make those businesses better? How can I save them from reinventing the wheel? Where should they outsource things? Where should they insource? Where should they use AI, and where shouldn't they? Just helping them make the right decisions for their businesses where they're at now.

Chris:
That sounds awesome. As a founder, that resonates a lot. There are a lot of things people learn the hard way, but they don't have to. They could use somebody like you to get a sense of what the broader market sees out there. You help a lot of people with tech stacks. What does a solid tech stack look like at five million dollars versus fifty million dollars?

Scott:
Yeah. It's crazy. I do a tech stack review as part of my standard process, and at about five million, businesses are at 30 different apps, spending between $2K and $10K per month. At 10 million, it's 40 or 50 apps, and they could be spending $10K to $30K a month. Some people are spending a lot less. And with AI, there's this huge variable in how much you're burning on tokens, like I am. At 50 million, you've kind of plateaued in some ways.

Without talking about specific product names, we can talk about certain areas. First, you need the right general ledger. In general, under 50 million, it's going to be Xero and/or QuickBooks Online. That's just the way it is. They can scale, as long as you only ask the general ledger to do its job. I don't ask my plumber to do my electrical. You ask the right tool to do the right thing. Don't try to track your purchase orders, costs, and quantities in it, even if you were on QuickBooks Desktop Enterprise for years. That's just not the right way.

Then you need an inventory management system or ERP, enterprise resource planning. That's the trusted source for purchase orders and all of your costs, including indirect costs. If you don't know what landed costs are: your buy cost is a dollar, but it takes nine more dollars until it's live in your warehouse, your 3PL, or Amazon. Your landed cost is $10. So if you're selling it at $7, you're not going to make it up on volume. It's the trusted source for your inventory quantities, whether that's in Amazon FBA or wherever else, and it's where every order comes in. Every EDI order should connect into the IMS, not the general ledger, in my opinion. Then it connects to all of your channels, pushes inventory quantities, and pushes limited information into your general ledger without corrupting it. I have a separate YouTube video I did with a partner on that. For 10 years at Catching Clouds, we wouldn't connect any of the inventory management systems to Xero. This was a few years ago. We had an air gap because they would corrupt the financials. So you have to understand the flow of those things.

Those two should absolutely be SaaS, software as a service. Deterministic, old-school code. You can read the data with AI and get feedback and intelligence from it, but they should be locked-in SaaS tools.

Next on my list, when I'm helping people as a fractional CTO to evolve their business and put in that solid foundation, so the business runs itself instead of running the founders and team members into the ground, so technology is helping you instead of driving you crazy: you need a data warehouse, data lake, whatever you call it, that pulls in all the data from the general ledger, inventory management, your marketing channels, and any unique data about your business. You don't want just a dashboard tool that gives you three marketing dashboards and TACoS and ROAS. I'm using a lot of acronyms, but if you're in the space, you know them. Once you have your data, you can ask questions about your business, and there are multiple ways to do that.

Once you have that foundation, then of course you'll have Google versus Microsoft. You'll have your payroll, which will be Gusto or Rippling or somebody else. You'll have your bill pay, which could be your bank. There are a bunch of great online banks; really get away from the old traditional banks. Then you'll have a few marketing tools and your customer service tool, and there's going to be some AI in each of them.

Then you can look at each area of your business. In sales: are we going to use AI or this tool? Are we going to use this CRM? Then marketing: what are our tools? We do a bunch of YouTube stuff, great, we just picked a tool to edit the YouTube videos, and we'll start doing more of that. You end up with a cluster of tools, then processes, and then the AI that wraps around them. Is it an AI agent doing a bunch of the work? Is there a human in the loop? You should always have a human in the loop somewhere. And how do those pieces all come together? That's where you end up oozing AI into things. You can have a big, grand plan to do everything with AI. Well, you can't do everything with AI, but it sure can supplement and optimize things. You have to have a solid foundation of accurate data first.

Chris:
That's a really important point, and one that gets missed a lot: the foundation, and how you're putting all that data together. What are you seeing out there in the mid-market for that data lake or data warehouse?

Scott:
I have maybe 20 different comparison matrices that I don't share publicly. I have an IMS one that's about 70 vendors and 400 fields deep. I have ones on banks, sales tax, EDI providers, data warehouses, dashboards, listing tools, and all these other things. When it comes to data warehouses, it's kind of a blend. In a lot of cases, like for general ledger or outsourced accounting, I'll take two SaaS tools and compare them. Most of the time when I'm recommending a data warehouse, I'm recommending a SaaS tool and a custom option. I have multiple custom data warehouse consultant partners that I vet.

A lot of times I run into people who have five different reporting tools: one that looks at their CRM sales data, another for their subscription data, another doing specific financial data, and so on. If you add up that whole budget, it's going to be two, three, five thousand dollars or more, and you could actually spend less than that on a data warehouse after you've paid to get it implemented.

For companies that don't have data expertise or really any technical resources, and have minimal time, which is usually the case, I recommend going month-to-month with a few key tools. I don't really have favourites. I usually have some to avoid, but research your options.

Now, if anyone's trying to figure out marketing attribution, that's black magic. Can I tell if this click on Instagram, that went here, that went there, that hit a Google ad, that landed on my Shopify site, and then they came back later and bought it on Amazon... can I track the dollar I paid for that click to this sale? No. Total black magic, because these other companies make it so complex. You can look at it financially: I'm spending this much on marketing in this area, and it's generating roughly this much in sales. Cool, there's an ROI there.

In general, I'd recommend getting a dashboard tool or a data warehouse tool that specializes in e-commerce. If you search with AI, you can find some. Pay month-to-month, pay two or three thousand dollars for them to set it up properly, use it for three to six months, and figure out what you want. Then you might define your top five KPIs: contribution margin for the total business, which can be in your general ledger; contribution margin per channel, meaning how much money am I actually generating to the bottom line; and contribution margin per SKU, which can be different per channel, because you'll make more money on every Shopify sale than on any marketplace sale.

Then you'll decide: if we just do AI, we're going to spend $600 or $800 in tokens, but this tool is $300 a month. So yes, AI is cool, but we'd actually spend more. In other cases you might say, we're a big Google shop. We'll use Google BigQuery to pull the data together, use AI to build it all, and constantly evolve it so we can always ask the questions we want. For the people who really get it, I still recommend you start with the right consultant and build a custom solution, whether it's on Google BigQuery, Azure, Snowflake, or another true data warehouse. You still need somebody to use existing tools instead of writing from scratch for what's called ETL, or ELT, I never remember which. Those are the connectors to Amazon, Shopify, Meta, Facebook, Google Ads, and everything else that pull the data together. You don't want to waste a bunch of time tinkering on the engine. You want it to actually be doing something.

What you're trying to get to is trusted data in your financials, trusted data in inventory management, and trusted data in the business, so you can ask questions that give you information to make decisions. You know where the numbers came from, you know you've got the right processes, and it supports the business. Then you can wake up in the morning and connect your AI to it.

I played with Duet this week. If you don't know, there's Duet, Dot, and Grok bot. These are three kind of personal agents that are always on. Mine kept telling me I had a conflict with this meeting because I have all day Friday blocked off, even though I booked this knowing I had nothing else planned. I had two or three different bots tell me I should contact Chris to reschedule.

But once you have it set up, you can have your morning briefing come up, or at lunch, and just ask one question: "Spock, how are we doing?" or whoever your alter ego is. My main Grok bot is Elrond. It's all Lord of the Rings themed.

Chris:
Is that right?

Scott:
All my agents. How's my business doing? What do I need to do today? What were the sales? What was my return percentage? What was my conversion rate? Did it change because I started this new ad campaign yesterday?

Chris:
So not Heimdall, not Gandalf?

Scott:
Well, we have Gandalf, but I keep building AI agents and killing them off. I had Gandalf the Grey, then I killed that one off and built Gandalf the White, and now I've killed that one off too and moved on to other things.

Chris:
So no MCU, no Heimdall, nothing Nordic?

Scott:
No. If you can see, I've got a little bit of other stuff and some space things, but in general my AI stuff tends to be Lord of the Rings.

But the approach is: what are you trying to do? If you want to sleep better at night, have detailed, accurate financials. That's somebody else's job, an outsourced e-commerce accounting firm (we've sold ours) taking care of your accounting. You're not an accounting firm, so don't waste your time doing it as a founder. Have somebody else do it, and that'll help you sleep better. Then, when you can ask questions about your business and understand the answers, it puts you in a different place. Once you have that foundation, you can leverage AI agents and other things and do amazing things on top of it. But if you haven't put the foundation in, AI is either making stuff up or giving you the best information it can from bad data.

Chris:
Yeah, we see this a lot in the EDI space, where somebody gets really excited: "Look what I've done. I've taken that one file and translated it. Isn't it amazing?" And the thing we always come back to is that EDI is deterministic. It's very explicit. LLMs are probabilistic. If you have an extra carriage return or an extra period, you're potentially talking about tens of thousands of dollars in chargebacks and dinged scorecards. It's a lot to risk for somebody who's very excited about their new project, and we see it all the time. There are certainly applications for AI, a ton of them, and we're working on many. But it's amazing to think it's going to be this magic wand. If it's making judgments on the wrong data, that's a big issue.

Scott:
The other challenge, when we talk about deterministic versus agentic, is that they're in this race to release new models. New models think in a different way and can give you different answers. You might try to lock down the model, but some of the older models people love are no longer available. A model release can give you amazing new things and more intelligence, but it looks at things differently.

Take demand planning in e-commerce. What's the velocity of our sales? What should we buy so we don't stock out? What shouldn't we overbuy, so we don't have cash locked up? You might have something that worked last season, whether it's your own spreadsheet or a tool. You're very seasonal, you're focused on Q4, Black Friday, Cyber Monday. It worked last year, but if you're using AI, the equations have changed, or the AI looks at the data differently. It could look at the same data and give you a completely different answer. Could it be better? Yes. But can you trust it if you're basing decisions on historical data? That's the real challenge. You just need to understand those things so you know what you're asking it.

If you spend enough time with AI, it's just like dealing with your own toddler. They're the most amazing thing, you love them to death, it's like magic, and then: why did you take a Sharpie to all the walls in the house?

Chris:
The CEO of Notion was talking about the difference in an interview. He said engineering is engineering, but AI is more like brewing beer. Sometimes you put in a bunch of ingredients and it comes out a certain way, and it's a pleasant surprise. Other times it's not. It really depends on what kind of surprises your business can handle and what kind of judgments you're making from them.

Scott:
The more you use AI, the better you get at finding the balance between locking it down to do certain things or not do certain things. If you don't know about negative prompts: don't use the word "synergy," don't do this, don't do that. For me, you've got to show me data in tables or bullets. If you give me big blocks of text, I'm literally not going to read it.

Chris:
I hear you.

Scott:
So you need to train it to work the way your brain works, to not do certain things, and to not be a sycophant. But you can't make it too mean, either. I have a bunch of prompts that challenge my biases, to make sure I'm not just doing something because I've always done it that way. But I made one that was just too mean.

Chris:
Bad boss, bad bot.

Scott:
Throughout a report, I'd ask it something and it would say, "Yeah, but consulting's dead." This was six months ago. I do consulting, I want to do it for the next 10 years, and I've done it for decades. And it's saying, "It's dead. If you don't write this AI thing, you'll never make another penny." It was just horrible. I finally said, "Stop. You're allowed to give me three bullets at the end of the report with some critical thinking, but it can't be totally negative. And the second sentence has to be a positive response on how to address the real threat you're pointing out."

Chris:
Nice. That's interesting, because I know you're a huge early adopter and practitioner of AI. You test all these things. I remember when we were first introduced, you mentioned your spreadsheets, where you keep everything, with different categories for your analysis of these spaces. Are you moving any of that over? Are you creating a virtual Scott we can ask questions of?

Scott:
Well, not quite. I haven't dumped my total consciousness in somewhere yet. I have a number of processes I've used for years, and I went back through my prompts and turned them into a prompt library. What I'm doing is moving from Google Sheets, where people give me the details I need about their business, themselves, and their stories, into a database. Then I've built a library that runs all the fields through a whole set of how I look at those things. It tells me specifically where to look and where the gaps are. Then I review it, human in the loop, make comments, and direct it. Then I run it through a set that generates cool infographics; you can dump the data into NotebookLM or Gemini Notebook or whatever they're calling it now. So I'm turning more and more of my work into those things. I'm focused on the survey and consulting CTO work first, but I'll be doing more.

I have AI projects in multiple tools: some in Grok, some in ChatGPT, most in Claude and Grok. I have one for IMS/ERP, one for data warehouses, and so on, and I have them self-learning. Probably every month I have each one review its own instructions and previous threads, pull things out, and update its prompts. It's recursive self-learning, RSL, if you've heard about it. How do I use those tools to analyze the way I work so they do a better job supporting me? Sometimes they're getting better, and sometimes worse in some ways. There are so many nuances to each one.

I use them adversarially for everything. Any code I write in Claude Code is reviewed by Grok and Codex. They each find different types of issues, and it keeps fixing them. It's the same as getting external advice. Some people ask one opinion, some people ask a dozen. And most people tend to trust the groups they're in. If you're in a mastermind of other e-commerce sellers, you're going to start there. In some cases it's great advice; in others, their business selling supplements is totally different from your business selling baby clothes. You're doing manufacturing and they're not. Some people say, "You trust that company? I'll go with that company," but it's horrible for their use case. Same with an online community, or whoever.

Unfortunately, there are a number of e-commerce sellers who don't know any other e-commerce sellers. None of their business partners or family members are entrepreneurs or in the same business. I'm constantly introducing e-commerce sellers to other e-commerce sellers, because you need peers. If you haven't found other e-commerce sellers, or even partners like Chris and myself who live in the space and know all the three-letter acronyms I've dropped more than usual here, find them. People like us are going to charge you money for certain things, but we're happy to give good advice. And you want to find those peers.

Chris:
That always drives me nuts, when people are so averse to jumping on the phone and talking. "Just email me the answer, email me a price tag." I understand not every conversation is going to go the way you want, and not everyone's going to provide value. But if someone's been in the industry for 20 or 30 years, that might be an opportunity to learn something, maybe even two things. It drives me nuts.

Scott:
Yeah. But entrepreneurship is amazing, and I'll do anything I can to support it. You're taking a risk with your kids' college fund, your home equity line, and a crazy amount of your time. You're trying to build generational wealth, or whatever your goals are. You don't want to work for anybody else again, so you're going to be an entrepreneur. Fair enough. And there's so much to learn. You become a lifelong learner. Entrepreneurs are diving into AI to see what they can get out of it, but everybody needs more perspective. My mom asks, "You work out of your office all the time, aren't you lonely?" And I'm like, "No, I talk to people all day long." Some days I have 10 meetings. I have relationships with my partners, like Chris, and with my clients. It becomes a community.

I started a tool in 2017. This was way before AI, way before vibe coding. I had the right idea and the right plan. I learned a ton, and I still use a lot of what I learned. But I could have a lovely, paid-off mountain home in Colorado for that money. I picked the wrong developer, a very nice guy, but I'm doing more myself every day now than that developer did in three months. And I picked the wrong platform, an early open-source RPA, robotic process automation. It was the right direction but the wrong thing at the time, and now it's all been consumed by AI. It was a distraction.

I saw a million-dollar fraud that the FBI got involved in because a company didn't implement an inventory management system and didn't have internal controls. They were going too fast and not taking care of fundamentals. They weren't creating purchase orders. They were just calling people up, ordering, and giving out their credit cards. Then somebody realized they never paid attention, so they just kept charging the credit card every month. We had a report where we could tell there was a hole in the business. Money was disappearing, and we couldn't find it. We thought there was a million dollars in lost Amazon returns, because they were going so fast. We dug into that data for a few months, and it was really hard because they didn't have an inventory management system.

If you're an e-commerce business, the sooner you get away from spreadsheets for purchase orders and inventory tracking, the better. If you're under one to three million dollars, there are a few tools out there under two or three hundred dollars a month that will get you off spreadsheets, give you great capabilities and much better data, and help with accounting. You can always reach out to me and we can have a conversation.

Chris:
Most of those early brands are running cash-based books, and I know you're a big fan of accrual. Maybe you can explain that.

Scott:
So here's the deal. In your business, you have your income: I sold $100,000 this month on Amazon, Shopify, wherever. Then you have your cost of goods sold, which is $10 per item. I sold a hundred, so that's a thousand dollars. That's the income statement, or P&L, profit and loss statement.

The next piece is your balance sheet, where you hold your assets and liabilities. If you have $50,000, $100,000, or a million dollars of inventory sitting in a warehouse, that's cash in another form. It's just not dollars in the bank, or in a safe, or buried in your backyard. You want to be able to track that. You also want to see your liabilities: we collected $50,000 of sales tax that we owe in three months. You need to know that, so you don't spend that $50,000 and not have it when the sales tax is due. Or, we owe this much on our Amex Plum card. Maybe you're on Amazon, and every time you get your settlement statement, you pay off your Plum card and buy more inventory. That's great. But I think every e-commerce small business should run its books on accrual. Then you see the value of your inventory. Say you have a million dollars in inventory, and $200,000 of cost of goods sold for what you sold and shipped this month. That comes out of the inventory number and moves over to your P&L, and it shifts back and forth, so you know where you stand.

The other thing: say you paid $12,000 for insurance. You can accrue $1,000 a month. So instead of "we lost $2,000 this month because I paid $12,000 for a whole year of insurance," it spreads the cost over the year, so you can see how you'd budget and balance things out.

Now, that's completely independent of how you file your taxes. Your CPA will tell you whether an LLC, S Corp, or C Corp is the best structure for your business and gives you the best tax advantage. You can file on a cash basis and run your books on accrual, which Xero, QBO, and all the general ledgers will do. If you run your books on accrual all year, it helps you make better decisions as an entrepreneur and business owner. Then it's about 30 minutes of work in January to convert your accrual books to cash so your CPA can file on a cash basis, if that's best for you. You can't just say, "I'm cash basis for my taxes," and stop there. That's independent of how you need to look at your financials to make the best decisions. And it's a combination of bank feeds and pulling the data in directly.

As soon as you're big enough to outsource your accounting, do it. It's the same as marriage: if you get to the point where you can afford to pay someone to clean your house once a month, it will change your marriage. That's a major milestone. That's one of the fights. Now, I've been married 30-plus years. The fight over the thermostat will not go away. I can't help with that one.

Chris:
You don't have one of those plastic lock boxes?

Scott:
No, no, no. We both have it on our phones and can mess with each other.

Chris:
That's awesome.

Scott:
But as soon as you can outsource your accounting to an e-commerce accountant, someone who specializes in it and is certified on A2X, you'll get better numbers. You'll save a ton of your own time, and they'll do it properly. And keep in mind, accounting isn't real-time. It's a daily, weekly, monthly thing. You can't trust your books for the prior month until an accountant says the books are closed. That means they've reconciled every account: your Stripe account and any special fees, your bank account, all the different accounts, and any inventory adjustments for lost, stolen, or damaged goods. It takes a number of hours to finish your books, and once that's done, you can trust your financials.

The nice thing about double-entry accounting is that it's about 600 years old. If it's done right, the balance sheet balances, and all the steps are taken, you can trust it. You can take it to the bank. You can take it to the government. More importantly, you can take it to your spouse and say, "You know how I was busy 50, 60, 80 hours a week last month? Here's the result." It could be a bad result or a good one, but at least it's the truth. It's a process, like blockchain, that works its way to the truth.

So hopefully that was a short enough answer. But it's a better way to look at your financials, so that you, as a non-finance, non-accountant person, can understand: Is my business profitable? How much money do I owe? And how much do I have in the bank, or in inventory I can sell for more than I paid, to keep the business going?

Chris:
Gotcha. Do you ever get involved with people who are evaluating whether to double down on their own warehouse and warehouse management system, or use a 3PL? Are those decisions you assist with?

Scott:
Yeah, with the high-level decision and then the tech stack to support it. About a decade ago I had a short list of good 3PLs and bad 3PLs, but it changes every day. If you're early on and still in your own garage or your parents' garage, the sooner you can get out of there, the better.

Probably the biggest factor in deciding between your own warehouse or a 3PL is how you want to finance. At a certain point it gets to be a challenge to pull money out of your business. If you start making real money, there are just challenges. So what are the considerations, depending on how you want to scale? If you're scaling internationally, you're going to want to look at a 3PL model. But if you want to maintain control, a lot of sellers will either buy a warehouse or buy a plot of land and build one. You put it into your family trust, the family owns it, and your business pays the rent and all the improvements. You're building equity for your family.

So that's one of the key decisions, but it's more of a personal choice. Do you want to control quality? Do you want to hire locally? A lot of people build a business not only to support their family and pay for food, vacations, and their kids' college, but because they want to be part of the community. You want to be a better boss than the bosses you've had. Or you have family members who can't get jobs anywhere else, and you might as well hire them, which is good and bad too.

Dealing with 3PLs can be a real challenge. Some are great, and then they get awful because they have a bunch of turnover and it's hard to keep them staffed. Dealing with them is one of the biggest challenges in implementing an inventory management system, because in my opinion it's not fully operational until all the data flows from all your channels to your IMS to your warehouse or 3PLs. All the routing, including all the EDI orders, comes into your IMS, and that's your order management control. That's how you avoid split shipments, where somebody orders five products, it ships from five warehouses, and you lose money on the order. If all of those products were in one warehouse, it goes to that one warehouse, in one box, and it's still profitable.

Chris:
And then what's the single version of the truth at that point? Is it what the 3PL has?

Scott:
Depending on the 3PL, you have to have an inventory management system so you have your own truth. It's trust but verify: you reconcile and compare. I had a seller who gave us a report of their inventory, and we said, "You had $50K of value disappear." He sold a single SKU, a board game. He was like, "Oh my God, what am I going to do?" So he ordered another year's worth of product. Six or eight weeks later, they found a cage with all of his product. He was able to sell through it all, but he could have spent all of his profit for the year on extra inventory.

If you don't have an inventory management system, you don't know that you bought a thousand units and sold 500, so there should be 500 left, or 500 minus five, because some will be lost, stolen, or damaged. Things are going fast, and that just happens. People who go straight from their Shopify site to a 3PL and just trust the 3PL's reports could be receiving the wrong amounts, or it could take days or weeks to receive the right amounts. Without solid financial numbers, it's hard to stay in business. You need an inventory management system you control, so you have your own truth and can compare it against theirs. Then you understand that a certain percentage of things are lost, stolen, damaged, or disappeared, and you can account for those properly and write them off in your financials.

Chris:
That's great advice. What's usually the sign they've hit an inflection point, where the entry-level shipping app, whatever that might be, isn't enough and they need to move to the next tier?

Scott:
If you can get through the basics, it's really about running your own statistics, surveying your team, and looking at costs. If you've optimized your credit situation, you're buying minimum order quantities from your vendor so they're not charging you shipping, you've redirected things to deal with tariffs, you work down those things and then look at your shipping. Do you charge for shipping? Do you do free shipping, or free shipping on orders over a hundred dollars? When you look at all of that, you'll realize you can't make the right call if you can't see your profitability per order, if your warehouse management system doesn't track that per order. So you need to track the data first.

Then you do the analysis, and AI is great at this. "We should be shipping all our stuff in mailers or this USPS box, because it's a quarter of the cost." There are some good consultants who'll help you analyze it, but you can also use AI to look at all those pieces. Have you nailed those fundamentals? You could change your bundling or packaging in a way that transforms the flow and makes it more efficient for the people moving things around and boxing. And then, can your warehouse management tool support it?

The idea is, if you're not dealing with a hundred other issues in other areas of your business, you can say, "This month or this quarter, we're focused on shipping efficiency." And one of the questions should be: do we like or hate our warehouse management tool? Take into account the opinions of the people you hire, even someone who's only been there a few days. If they think it's the hardest tool to use, that matters. If you look at an app and the UI makes your brain hurt, don't pick it. It might be half the price of another one, but don't do it.

Then build your own requirements, iterating with AI or with a consultant if you need to. "We want it to do this, we want it to do that, it's got to have a direct integration with DHL because we ship into Canada all the time." Go through your requirements, like in all of my processes: What are the good things you want to happen? What are the bad things you don't? You probably have a huge list of bad things that have happened with bad shipments, returns, and scams. And what's the optimal world? If you can define those things, you can give them to AI, or a consultant like me, to narrow it down and pick two or three. Do not look at 12 tools.

And be aware that a lot of the review sites out there are pay-for-play. A tool will get a bunch of good reviews, and that's not real. Those aren't real customers.

The idea is to focus on a specific area of the business, really optimize it, and go through this straightforward process, rather than "Hey, there's this certain tool." That said, if you get over 30,000 or 40,000 orders per month, there are some very good WMS tools you should step up to, and they'll transform things. If you're investing in your own warehouse, you might even decide to offer some of the space to your friends and become a mini 3PL. "We need 10,000 square feet now, but I'm going to build a 50,000-square-foot warehouse. We can play basketball or indoor hockey in it with the family for a while." Then: "Wait, I have a friend who hates his 3PL. I can just do that. I'll hire one other person and get a tool." That's another reason to step up to a WMS that lets you keep track of Acme and Beta and Delta and whatever other companies are in there. Then you get economies of scale and help pay for the lease or the mortgage on the warehouse you bought.

Chris:
That's wild. That probably explains the genesis of so many of the thousands of 3PLs that have cropped up in the last ten years.

Scott:
Yeah. For a while, everybody thought Amazon FBA was going to kill off 3PLs, and it challenged them for about a year, even with AWD, which is kind of a staging warehouse for Amazon FBA. But what it actually did was shake the market up, and then it exploded. There's just so much more out there now.

Chris:
Right, the Jevons paradox in play. That's wild.

So let's talk about EDI for a minute. A client comes to you and says, "Geez, we're moving into retail. I've got a lot of concerns. I want to make sure I'm set up to win." What's the process you walk them through, and what questions and concerns do you help them with as they make that decision?

Scott:
One of the first is: are they aware of what they're getting themselves into? Educate yourself, because a lot of these retailers have chargebacks and marketing fees, and they can send your inventory back. You have to understand what you're going into. Use AI to rip the contract to shreds and shake out all the potential scenarios.

Another is that I bet the contract will mention one of Chris's major competitors, but more than likely it doesn't require it. A lot of people assume it's required because it's in the contract, since that's easy for the retailer. Unless it's actually required, there are better options with better support or better pricing, like Surpass.

The other piece is friction. Can you dip your toe in? There's a lot of overhead. For the first few months, if it's a handful of orders and there's no volume, you can just use the web front end and have somebody download and enter orders. There's a small piece you don't necessarily have to automate. Some people jump ahead and say, "50% of my business is going to be B2B," and spend $20K up front, and it takes them four years to get to the volume where they needed that tool. So you want some balance. But as soon as you're getting any real numbers, probably even just a couple of orders a day, you want to implement the automation. And it really should go from EDI into an inventory management system, so every order is tracked back and forth and you're not doing double data entry. You just want to remove friction.

There are also things on the IMS side, like using virtual warehouses to set inventory aside so it's always available for your EDI and retail customers. Make sure you've got the right inventory quantities, so you don't oversell, or sell a bunch online in a big flash sale and then can't meet the demand on your contracts.

A lot of it is just going in with open eyes and reviewing the contract. But as soon as you hit any real volume, you really want to implement EDI and automate it to an IMS. Then it can grow and scale, and you can adapt. And there are a bazillion weird, crazy nuances. EDI is going to be around as long as cockroaches. It's ancient technology, almost like Morse code or stone tablets. But all of these retailers just aren't willing to invest, or they've already invested in these mainframe systems that talk EDI.

Chris:
I don't think anybody's really come up with a decent enough replacement. It was supposed to be XML, which has quintuple the payloads and requires a lot more. But the same thing happens when you try to translate a literal business process into data. Everyone has a different business process. Even within their own business, they have different processes for how they handle different goods, and that all manifests itself there.

When someone says, "I'm only getting a few orders a month," or "a dozen orders a month," they think about the order, but they don't think about the acknowledgement, the shipping notice, the invoice, the warehouse shipping reconciliation. Take that times six, multiply it by three or four minutes each, and at the end of the day they're eating up thirty hours a month without even realizing it, because it was one of those boil-the-frog situations.

Scott:
And then there's the accounting side, being able to reconcile all those things if you don't have good data. You've got some retailers where it's, "Okay, they'll give me an order twice a week, it all goes to one distribution center, easy peasy." Then you get your next one, and they expect you to ship to 40 different stores three times a week. Before, you'd get an order and it shipped to one or two places for a consumer. Very straightforward. Then you sign this up without thinking about it. Let alone all the nuances people don't know about: you might have to change your packaging, change your labeling, they expect labels in a different place, and it's different from the last retailer. So you need different external boxes or packaging for different retailers. It compounds. If you don't have accurate data, you just can't keep up once you're dealing with all those other realities. It's about asking all the potential questions so you know what you're getting into.

What I've seen, in the right market space, is everyone saying, "Cool, I got Whole Foods," or whoever. You really need to analyze whether it's more profitable, because each of your Shopify direct orders might generate more contribution margin than 20 EDI orders, between the marketing credits, wanting to be on an end cap, all those things, let alone your time negotiating the contract. Now, keep in mind it can be slow and steady cash flow, and you're getting your product out there and building awareness. People see it in the store and then buy it online. Those things build, and it's a good strategy. You just have to know what you're getting into.

Chris:
You're right, a lot of people don't know what they're getting into. They think they're going to be wildly successful, and they don't realize how capital-intensive it is, because you've got to wag the dog in a lot of instances. Then there's the infrastructure, and then there are contracts written by Lucifer on the retail side that you've got to watch out for.

Scott:
Well, net 120, and then they still string you out another 30 or 60 days.

Chris:
Less charges and allowances.

Scott:
Oh yeah. Less charges and allowances, which always happen. And then, wait, I don't trust my financials. We helped somebody clean up their books. They were selling pet products on Chewy and PetSmart, and when we cleaned up the accounting, they realized no one had configured Chewy to pay out. They were selling lots of stuff on Chewy, and the money was just sitting there because they hadn't set it to auto-disburse every day or every week.

Chris:
That's a nice little loan.

Scott:
Yeah. And all of these companies... Amazon does it deliberately. It does everything it can to hold your money as long as it can, because it uses it to fund everything else. You're loaning them money at no interest, whether it's for hours, days, or weeks in their settlement statements. They've really got a deal there, where they're using your money to do a whole bunch of other stuff to make more money, and they're able to get more value out of your money than you can.

Chris:
Right. It's almost like they have people whose sole job is to figure that out. Because they do.

Let's talk about the brand that's already there. They're using an EDI system, they've grown with it, and obviously they've done well by it commercially. How do they determine this isn't the right direction anymore and they need to go in a separate direction? What are the signals you help people look for?

Scott:
Coming back to basics: you have to have accurate numbers. You have to be able to compare your per-channel profitability and volume, and how that impacts your cash, versus per product and per order. For any business, you really need to put your business on a cycle and pick big focus areas each quarter. "This quarter we're reviewing financials, our lending, our bank." "This quarter is all about our vendors, our products, and the factories." And another quarter, especially in this day and age of AI, is technology: let's re-evaluate and challenge it.

One thing most people do is set up a tool and never revisit it. EDI kind of gets set and does what it does, but with a lot of other tools, people get a larger IMS or something else, never update it or use the new features, and then buy three or four other apps. I'll go into a $5 million or $10 million company and find $1,000 or $2,000 of tools that are totally redundant, with better versions already in their existing tools.

So you want to look at what you're changing and not changing. Once you're on a general ledger, Xero or QuickBooks, I tend to recommend you don't change it. Don't put your accurate financials at risk. If you have an IMS you're happy with that's more than 70% working for your business, focus on optimizing it. Then look at the other areas where you're connecting data, like EDI. You absolutely want to look at that, both the cost and everything else. Should you be paying annually? Is it a pain in the butt? How's their support? What are you able to do? What were the last few experiences? If you can stop long enough to look back, or you've documented some sort of timeline, you might realize, "We had issues every three weeks." You can use AI to analyze it, too: how many support tickets did I send to XYZ company?

Take a step back: we made these decisions in the past, and this is our once-a-year time to challenge them. Not three weeks later, but the next year. Where can we build improvements? Where can we remove even more friction? If you've got the general ledger and IMS/ERP locked down, that's the time to look at the other pieces around them, including EDI, and ask: how do I optimize that? How do I optimize the integrations to my 3PL warehouses? Or am I changing 3PLs?

Put it on a schedule. The first month, you ask all the questions and analyze. Then you get some external advice. That's when you should really challenge those things, and challenge everything. Changing a general ledger is a big deal. If you switch from Xero to QBO just because you changed accountants or didn't trust your financials, you have to change the whole process. But EDI is one of those tools where you can build efficiency. With any company where there's friction, where you get no support or horrible support, or it goes down, if you don't stop and look back, you won't realize how often it was down, or that you opened 42 support tickets, because you're so used to it. That's one of the reasons to change.

If I saw that data or asked those questions, I would absolutely have somebody talk to someone like Surpass. They're going to make the transition to their platform easier than dealing with all those support tickets, or even setting up a single new retailer. Think about it: "Oh my God, we couldn't start with this one retailer because it took an extra six weeks, and we lost that potential revenue, or we sat on that much inventory for that much longer, which impacted our cash flow, which means we didn't kick out distributions the way we wanted, and our wives were all mad at us." Or spouses. If you can step back and look at all those things, that's when a company like Surpass makes sense. A company that's smaller, more focused on the vendors, founder-owned, not driven by VC money, not changing its mind and buying and selling random companies, not just trying to eke SaaS money out of you, that's really part of the community. Those are the kinds of changes you can make.

Chris:
That's great. Thank you so much for that, I really appreciate it. It's true: when you're motivated by working with entrepreneurs, like you are and like I am, what a difference that makes in your general happiness. Instead of just trying to make all the numbers tick, which is also important to growing a business, when there's a connection and you can really help somebody improve their business and their operating environment, it means the world. You can really see those changes take place. So yeah, that's really cool.

You want to go speed round into sci-fi?

Scott:
I can talk about sci-fi if you want.

Chris:
How about an author that most regular, non-sci-fi guys would get into?

Scott:
David Weber, the Honor Harrington series. If you like space battles, there are lots of books, something like 30 in the series. An amazing character, a female heroine, cool techie space battles, but lots of other interactions too. It's really, really well done. That's my favourite author. Now, yes, I do read Lord of the Rings every few years, but David Weber... he has a number of other series too.

I've been reading Dungeon Crawler Carl, like so many other people, and I'm really looking forward to seeing it as a movie or an AI-generated world or whatever. I try not to read it before bed, because it's challenge after challenge. It's a combination of Hitchhiker's Guide to the Galaxy, D&D, and reality TV, this huge combination of things. It's really interesting, the characters are hilarious, and it'll make you laugh, but it's just too much intensity.

Chris:
That's awesome. Well, that's two, and I'm a big believer in the list of three. So you got another one in there?

Scott:
Old-school Anne McCaffrey and the Dragonriders of Pern. Just interesting stories, a little bit of sci-fi, some fantasy.

Chris:
I always get stuck with the same old ones, the most recommended, and never branch out. So those are some I haven't heard of. I'll be sure to get into them.

Scott:
Okay, cool.

Chris:
Awesome. Thank you so much. I really appreciate your time today, Scott. This has been great.

Let's connect!

Let's connect!

Let's connect!