Noël Perry talks about the intricacies of forecasting rates, especially on the individual lane level, and focuses on pricing strategies currently used by transportation providers, and where they can be made more efficient with predictive analytics.
Anticipating the launch of Rate Forecasting, the new week-by-week, lane-by-lane forecast from Truckstop.com and FTR, we sat down with Truckstop.com and FTR Chief Economist Noël Perry to talk about rates. This the fourth in a series of articles about the topic.
Nick: Hey Noel, we’ve been talking a lot about the different forces that push on rates, from seasonality, to events like hurricanes and ELDs, to capacity utilization—even to the psychological stuff that gives truckers the courage to ask for the higher rate. Forecasting a rate, especially specific to a lane, sounds complicated!
Noël: Well, there are two issues that are nicely handled by Rate Forecasting by Truckstop.com. The first is that spot rates have a strong seasonal pattern, so in any given week when a rate moves, it might not be because market conditions change in a sustained way; it may simply be that it’s a seasonal move. The FTR Truckstop.com forecasting product corrects for seasonality, so we’re able to isolate how much of a move is seasonal and how much of it is lasting.
Another thing that makes Rate Forecasting all the more valuable is it turns out that in any economic series in any business data, there is a strong random aspect. So if you’re measuring rates (we’ll say from Boise to Portland), some weeks they’ll be up by 15%, and some weeks they’ll be down by 20%, even if the average is flat. One of the things our work does is to quantify that randomness and correct for it in the forecast.
Let me give you an example of how that works, with that Boise to Portland example. Let’s say we think that, because of ELDs, rates are going to be up 10% a year from now. We take the latest rate from Boise to Portland—imagine it’s $2.20 this week. Well, if you’re not making these corrections for randomness or seasonality, you’d simply think that a year from now, it’ll be 10% higher, so add 22 cents.” So now you’re planning around $2.42, but what we haven’t corrected for is whether or not that rate this week has got a lot of randomness in it. Maybe the average rate over the last 3 or 4 months from Boise to Portland corrected for seasonality is at $2.01; so it would be incorrect to apply that growth factor to a peak that’s going to fall. In the same way, if this was a bad week and the rate was $1.80, you wouldn’t just add 10% to that.
What we do with Rate Forecasting is we correct that back to the underlying trend at $2.01, and we apply the 10% to that. So the 1st week it’s a little higher, the 2nd week a little lower, and finally it gets back to where it should be—so we can say with fair confidence that a year from now, the rate’s going to be 10% higher than the actual average rate of $2.01, which is $2.21.
It’s important to look at this data, understand its characteristics, and include that when thinking about the future. When people do it intuitively, as they’ve had to do in the past, they don’t get those corrections right most of the time. I mean every forecast is wrong, what we’re saying here is that if you do it by the seat of your pants, it’s going to be a lot more wrong than if you do it using some discipline.
Nick: Yeah, so you may be at a random peak for that lane. Plus you heard that this lane is in a seasonal peak. Plus you’ve heard that there’s economic factors. From this, you might just guess we need to put another 10% on top of it, but you didn’t know that you’re already at a random high. In fact, that Boise to Portland lane may never even show up at $2.01 on the spot market, the true going average, because it is so random day to day. You need rigorous statistical analysis even just to know the base rate to forecast from.
Noël: Yeah, and the same way you might compare a year over year rate trend (where seasonality is corrected out because it’s the same season a year from now); but if you want to know what’s going to happen in January compared to what’s happening today, you have to understand January and February or seasonal lows. If you want to know what’s going to happen in March or June, you have to add, because those are peaks.
Now, what this does is open up a whole new world of planning potential for fleets, 3PLs, and their customers, because right now they just plan one week out, and they say, “Oh, the number’s $2.10 this week; I’ll make sure I get $2.10 next week.” Well, what Rate Forecasting allows that broker, shipper, or carrier to say is, “Wait a minute. If I want to make my plans for the fall shipping season, I just can’t go by what what’s happening in August.” If I’m a fleet thinking about adding trucks or if I’m a customer thinking about having a big promotion, how I price that promotion has to include the supply chain cost. So if we think there’s going to be a big fall peak price effect, I need to include that. Well, we can do that now. In the in the past they were guessing at it, but now we have some data.
Nick: It’s a bit like a retailer moving luxury items out and survival supplies in ahead of a hurricane, to hundreds of millions of dollars of impact.
Noël: It’s the same thing. Any retailer or manufacturer, as they plan for the next season, knows the supply chain presents a massive effect on that.
Nick: So a product like Rate Forecasting should have massive effects on anyone from a small fleet to a massive supply chain as they plan their upcoming revenue and/or cost respectively.
Get a grip on your bottom line with Rate Forecasting, the groundbreaking, lane-by-lane rate forecast which projects spot market rates for each of the upcoming 52 weeks from leaders Truckstop.com and FTR.
Disclaimer: This post includes forecasts, projections and other predictive statements that represent Truckstop.com’s assumptions and expectations in light of currently available information. These forecasts, etc., are based on industry trends, circumstances involving clients, and other factors. They involve risks, variables and uncertainties. Actual performance results may differ from those projected in this publication. Consequently, no guarantee is presented or implied as to the accuracy of specific forecasts, projections or predictive statements contained herein.