Hey, I'm Patrick Escolas with another Banyan Tire Tracks podcast. I have with me Matt Silver of Greenscreens.ai. How you doing?

I'm great. Thanks for having me here. 

Hey, so one of the big things that we have here at Connect23 is we have Greenscreens here. We have a partnership with Banyan. We also have the CEO talking, but let's not let that – he doesn't have as big of a title, but he has just as much of the knowledge. What is Greenscreens? Start there. 

Yes. At its core, Greenscreens is a dynamic pricing platform to benefit the brokerage community, as we've talked about. 

This is a full truckload, correct?

Full truckload, yes. We've talked a little bit about it in some of our sessions but shipper product coming soon. Want to make sure we land on that. Today, we're selling just to the brokerage community, helping them find dynamic pricing for full truckload van, flatbed, reefer today short-term. We have some development underway in the next couple months. We'll have a long-term predictive platform. So be able to predict up to a year out. 

If you're a shipper, close your ears for a few months. But if you're a 3PL or a broker, listen up. When you say short-term, what does that mean, just spot? Or where's your space?

Absolutely. We definitely play in the spot market. We predict up to two weeks into the future right now, which for spot market freight, most of that moves within 36 to 48-hour window. 

That's just freight. You couldn't tell me if the Guardians are going to make the postseason or not?

Not yet. I mean, I can show you my personal betting app and make some predictions. But you probably don't want to go anywhere I bet. 

No, no. It's all right. Yes. I'll take your word on that one. No, but tell me more. With that, and I didn't mean to cut you off there. Yes, two weeks out, and this is with freight. Are there any caveats with that? What are you guys seeing with that, with the data you're able to pull back there?

We're seeing a tremendous amount, both between our team, as well as customer demand. We're learning more from it. We're making more astute predictions. We're refining the models that we're using. Kind of backing up a little bit on it, we produce for our customers two different rates. One is the market at large rate. 

That's everything that's going on. 

That’s what everybody is doing. Okay. We see from a lot of companies in the space or companies that are trying to create pricing products that they have 5 or 10 billion dollars’ worth of data in the trailing 12 months. We are sitting on 20 billion in the trailing 12 months, and we're still very new to this market. We’ve been around for two years. We're using that. It's anonymized. It's consorted across all of our customer base. We're setting the baseline of what's happening in the market today and into that short-term future as well. Then on top of that and going a step further for pricing, we built out a machine-learning model specific to each and every customer. 

Oh, okay. Got you. It’s specific to that client or customer or that broker or 3PL. 

Exactly, right. 

And their data. Okay. 

So all the nuances of their customer mix, the depth of their capacity network, all the shipment characteristics, right? Up to 38 different characteristics on every shipment are analyzed, and we use that to tailor our predictions. All the customers on our platform, it's like my favorite stat in the world, single-digit variance between our predicted rate and their actual booked rate. 

What's the single digit? What's the single digit?

Blended, it's like 6.5%. 

Okay, fair enough. Yes. 

I’ve seen better. 

It's not 9.99. 

No, no, no. 

No, no, no. That's really getting a no. That sounds awesome. With that, is that data and that does equipment type come into that? If I choose reefer, is it going to give me a reefer-specific versus flatbed?

100%. Yes. It's actually really important to make sure that we get the equipment type specific on top of the $20 billion that we have kept as clean data and we use for our models today. Obviously, that grows as our customers grow. But we've actually thrown out about 35% of the data we receive. 

I was going to say you said clean data. What does dirty data look like?

It looks like the data we get all the time. 

That's what I send you in emails. Yes. No.

Yes, absolutely. We work with our partners like Banyan to set parameters for what clean data looks like ahead of time, so the integration flows smoother. But part of that also is scrubbing the data that we get on a continuous basis. It could be something as simple as, and I'm not accusing brokers of anything. I used to be in this world, but maybe you build a load under one customer because your preferred customer is out of credit. Then when the shipment moves, you bump it over to the appropriate customer. Sometimes, you fat-finger the numbers, right? 13,000 pounds could become 31,000 pounds. When we predict, we're baselining all the data we've received. 

It’s not relevant. 

Exactly. 

Right. You don't want to build on that because –

Yes. Well, we get –

Trash in, trash out kind of a thing. 

Exactly. I actually love that phrase because I use it. But we reconcile our predictions against the actual booking, which is pretty unique. So all the data comes back to us as the actual booking, and we use the variance between those to retrain our models and redeploy them. 

Awesome. I get Greenscreens hooked up. I'm a 3PL broker. You give me a price. I just take a leap of faith and assume that's good. What else? What do you back that up with? What's there for me to use that number?

Yes. I love leap of faith, first off, because we will be right. 

Okay. Oh, you heard it here first. They’re right, just like my wife. Always.

I know it's very bold. We were going to name the product after your wife, but the name was taken. 

Hear that, Spenser. You're famous. 

We're trying to put everybody on the map. Honestly, what we're publishing, we prove it's right. Of course, there are times it's wrong, right? Data is fluid. It's not absolute. So these models are going to change over time based on all the factors of the industry. If we predict it on a Tuesday, you move it on a Wednesday. Certainly, that's going to change something. Even throughout the time of day as capacity enters and exits a market, right? That's the easiest way to think of changes, but we learn from it. We continue to make changes. We redeploy the models. Our goal is to get all the customers as specific as possible in the predictions. We continuously test new data sets as well. 

So you can imagine in our position, a lot of people come to us saying, “Hey, we have data around dwell time, weather, car accidents, freeway closures, wind patterns, dust in the air, pollution.” I think we've sampled all of it at this point. We run it through our data team to see does it bring us closer to 0% error? 

Tell me. How is that – I want to know how the dust – the dust is the important one, right?

It makes me sneeze. I don't think it changes the cost of freight. 

When sneezing while driving, there's more accidents, and the freight – no. I'm not as dumb about all this as I look or sound because I've heard a bit about what you said. With the numbers that the market and then the target that you can get from there, those numbers that you're predicting, whether that's all around or just client-specific from their data, you've got confidence and accuracy. 

Correct. 

What are those, and how are they different?

Yes. I really wish we had maybe used different names for some of these things. 

Yes, right. It sounds like me saying the same thing on an old test, where it's like, “Oh, yes.” No. Define accuracy. 

I'm confident it's accurate. 

Exactly, yes. 

Or it's accurately confident. We measure both things. On our platform, we talk about confidence, and we represent it in both the market rate, as well as the target buy rate that we published to our customers. What the market is doing and the confidence around that, what the customer-specific price is to go buy, the capacity and confidence around that. Confidence for our company in the way we represented it is effectively how hard do you have to work to go find the truck, right? We're giving you the price. We know it's accurate. I'll dive a little more on accuracy in a second here, but confidence around it is are you going to be able to go past or into your past carriers?

Confidence might be scarcity of the lane. 

It could be. 

Yes. How hard you'd have to work to get what that number is accurate for. 

It’s absolutely how hard you have to work, and it is reflective of capacity, scarcity, or abundance. 

Okay, got you. 

Your individual brokerage might operate different than the market at large because of carrier relationships. 

Understood. 

Go to market strategy tools at your helm, right? If you have a lot of digital freight matching products, right? You guys have great integrations with some of those players. If you're using those tools effectively, you probably don't have to post and pray, right? You don't have to dial for diesel or do any of the fun stuff that we're known for. 

I love those clichés.

I've booked a lot of loads in my day. I got something for everything. 

I like it. 

But counter to that, accuracy, right? No other pricing product in the market has a report card or an accuracy report. 

Really?

We publish ours live on the website. All day long, every customer can access it. Everybody –

You're self-auditing every day just to [inaudible 00:09:02]. Within that, this is a – have you had a day where you’re like, “Man, we're at 60%. It was a bad day because we had bad data.” Or what does that look like? 

No. 

Okay. 

Bad data gets kicked out. 

Right. That's fair. That's fair. 

The ML that we're using learns and retrains itself, and it continues to adapt. 

Would it just be if there's not enough good data then?

Yes. It won't regress necessarily. It'll respond to inputs. That’s why we're super cautious about what enters the data set because it's –

You don't want to throw off the whole thing and have it learning and we taught it wrong. 

Exactly right. 

It's really good at it, but it's on a flawed foundation. 

Absolutely. If there was some catastrophic event and all the tanker moves for a brokerage were classified as van and entered the data set, they would actually be so outside of market bounds that we put manual labor into checking it out and seeing what's going on before we let it influence any of the models. 

Okay. You've got all this data, and this is all truckload. You've had to have seen some trends, right? What is the current trend, and where is that going? Because outside of being able just to give a price, you've probably gotten a little bit of a road map of what's ahead or at least where we've come from. 

100%. Yes. Well, certainly where we've come from, right? Everybody in the industry is super reflective of how good things were for a little bit. It feels like –

Your VP still wants you to get all of that profit. They don't care that the market's bad.

Just 500% of your growth. That’s what you need.

Get it. Just find it. Make it happen. Yes. 

We’re in agreement with a lot of what industry experts are saying right now, which is we've more normalized then completely falling off the cliff. Things are maintaining, if not starting to rebuild, right now. For truckload rates, we're going to start seeing them increase. Capacity demand is getting a little tighter. So that's going to cause the abundance of equipment out there to dry up a little bit. 

There was a session earlier today that you guys had that was really, really great, where the small carriers that did suffer some of the shortcomings of economic downfall, they didn't exit. They just transferred that ability to move freight into larger corporations, and that's where capacity goes. Those trucks, obviously, come at a little bit of a premium, which is going to drive pricing up. 

Right. Do we expect things to keep moving in the same path forward? I mean, you're only a wizard for two weeks. What do you see six months out?

Yes. I think if we look back on some of the trends, right? We see that slow crawl up into the right, right? We're not quite a hockey stick yet, but things are starting to look better. I think when we talk in group settings, you get a lot of pessimistic feelings from groups. Every brokerage we talk to one-on-one. Every broker in this industry we talk to one-on-one. 

Isn’t every broker cynical by default? 

Everybody is super optimistic when we talk to them. 

Oh, really? 

I think we've had this kind of lull, right? A lot of people are using that downtime to reflect on where am I spending as a business, right? Am I inflated in my labor spend? Am I inflated in perks, right? Maybe not a new foosball table, right? Maybe less kegs of beer in the office. Let's focus on tools. 

No. Keep those. 

Some tooling that makes us more efficient as an organization. 

Yes. No. I like that, and it's great. Outside of what we've talked about, so you guys, I'm sure you have a standalone product. What's great about our partnership with Banyan?

Yes. Fewer mouse clicks, right? It’s easier for the brokerage to reach the answers that they need, right? That's what you guys have done a great job with. As somebody on the partnership side for the last five years of my career, it's great when the workflow feels native. It's super easy for the brokerage to capitalize from. Whether it's a sales rep or a carrier sales rep, they enter the information in their TMS. Immediately, at their fingertips, they have answers. 

It's right there. 

They know what it is going to cost them to acquire capacity. If they would like guidance on pricing to their customer, that's provided as well. Some of it inside the Banyan TMS. Some of it clickable out to RUI. We have a lot of – we can surface tribal knowledge and provide it as a coach to new staff. 

You bring up tribal knowledge, and I think that's a great point because for all those that have tribal knowledge, they're going to probably continue to rely on that. But I think a lot more people are – there's a lot less at each organization. People decide to retire, you go a different company, and then you left. You're telling a 24-year-old to go get these prices out, and they're going, “I don't understand the spreadsheet. I don't know who that guy talked to. How do I figure out what to get this pricing from?” I can see your tool as a great way to get them to a spot where they could be effective a lot faster. 

Yes. Tribal knowledge, I think, helps some of us, right? When I entered this industry, I was a young freight broker. 

Look at you now with your wise and wizardry.

I got the gray. I'm on my way to your haircut as well. But there's also –

It’s not my choice. 

As we go through business cycles, right? Capacity enters and exits markets. We go through peaks and valleys. Some of us forget what happened in a specific business cycle and how we overcame a problem. We might have that tribal knowledge rattling in the back of our minds. I usually think of it as like 3am I shoot out of bed. I'm like, “Billy's trucking company, right? He moved this for 1,200 bucks that one time, right? He was begging me for more.” 

Right. But I didn't have anything because this is a once-every-seven-years type of load. 

Exactly. We forget about those things. Embedded in our solution is something we call the negotiation coach. It’s awesome for young staff. It's also great for experienced vets because it's going to surface from our client's own data some of the tribal knowledge. Who's called it, what's the best and worst price you've paid for it, but also what are the factors that might influence the transit time? Is there a blitz week? Is there a federal holiday, something like that that could impact, and you want to be aware of it when you go to market? 

Yes. I also see kind of almost a soft value here in that this is going to – as somebody, a brokerage brings on new agents or new staff, is this a way to kind of ramp them up a little faster?

Absolutely. 

Have you seen that, or is this more of a hypothetical?

We have felt it. It's actually a fun thing that our CEO likes to survey the crowd about how long does it take to get good at pricing. 

Okay. What's the average?

I think most –

Without Greenscreens. 

Yes. When we blindly survey, it's anywhere between six months and a year is what everybody will just blindly say. 

[inaudible 00:15:46] fast. You got go through a whole year –

You get comfortable with some seasonality and some nuance to it. I will say in my entire career, we probably never got good at it, right? We had some gut feels. We would get faster at going to the tools. Who do we call in what region to check on some spot rates? Ultimately, it comes down to if you have a tool at your fingertips that can ping an API and bring you the answer in less than a second, that’s the best way. 

It's not about you knowing the pricing. It’s about having to go find. 

Yes. Look, I don't know how to do much of anything, but I know how to use Google very, very well. 

There you go. Put it on the resume, master Googler. No. I think that with some of the load board integrations we have with Trucker Tools, Truckstop, and then the existing stuff, whether you've got it a matrix or you're doing a spot quote that it's really more of an entire ecosystem within that truckload brokerage with Banyan now. 

Absolutely. 

Greenscreen can be used by itself. I think, really, it's better to bring into your entire process. 

I've tried to coin the phrase better together. 

Hey, I think they used that one a while ago. 

I saw it on the screen today. I love it. But really not just through our integrated partners on the TMS side like you guys. But we do have integration work going on with Truckstop. We're bringing in other products to our ecosystem as well. 

Right to the source. Yes. 

Just trying to string the most efficient transaction possible together. 

Okay. Hey, I've learned a lot. I learned a lot before we started talking from you. I keep getting – I feel like I'm still just scratching the surface. But one thing I always like to give anyone is what's your message either to any of the Banyan users, any of the listeners, anything, whether it's about you and how to get a beard that looks that nice or about Greenscreens? Here's my platform. Use it for a few minutes. 

Yes. I'll have a lot of advice about everything. First and foremost, buy green shoes. 

Yes. Can you show those? Lift it real high. 

I don’t know if I [inaudible 00:17:40]. 

You will hurt yourself. Please don't do that. 

You're going to see me fall out the screen on that one.

We're not insured for that. Yes. No. 

But with the platform, come check it out. Our team is at every event. I always like to say remind my sales team that they owe you a beer. 

They got the Hawaiian shirts. Look for the Hawaiian shirts. 

Absolutely. That's the next point. We're always easy to spot. Either find the beard or find the Hawaiian shirts. Our team is out there. 

I love it. I love it. Matt, thank you so much, man. I appreciate the time. 

Thank you. 

Thank you for everybody listening and watching. This has been another Banyan episode of Tire Tracks, and stay tuned for the next one. Thanks.