This week we continue our fulfillment conversation with Cliff, focusing on Amazon's infamous "Promise Date".
Transcript from Video
Kimberly:
Hey, welcome back. We're here again with Clifford and we're continuing our conversation about ship date, arrival date and promise date. Clifford, we had some really great information,
Clifford:
Kim. It was awesome. We geeked out so hard on this that we ended up going for over an hour on this topic. And I think we could probably go for even longer. So there's so much meat to chew on here. Let's just get back into it.
Kimberly:
Okay, so now let's get into promise date because we've kind of talked about arrival date and how we get there and optimizing networks. But now this is the big commitment I have promised you that you're going to get this on Tuesday and one of the interesting things that's out there is that at Amazon, if it arrived on Monday and I promised it to you on Tuesday, that's still counts against you. So arriving early isn't necessarily a good thing. So you want to talk about that a little bit because that was always a metric that surprised me at Amazon. Where we were like, Oh, but 10% of our orders arrived before promise date. We need to work on that. So let's talk about that a little bit.
Clifford:
Yeah. You know, early is just as bad because, you made a promise to the customer and the customer has an expectation. If the expectation is Tuesday, then that's fine. Now there's nothing wrong with delighting the customer. The challenge with it is the inefficiencies it's creating in your network. Yup. Right? So you have picked, packed and shipped something that could have waited another day or you have paid for, you've upgraded the level of service or something. Your whole internal fulfillment experience in a way that now is, it didn't really benefit the customer because you promise them something and they were happy with that promise because they made the purchase. Now you're just costing yourself money.
Kimberly:
Yup. Yeah. And I think that's super interesting to think about. And the way that I lay this out for clients sometimes is to think about it as a marathon. Right? So if you ran your first six miles much faster than you needed to to make your overall time, you may be hurting your last 20 miles or 15 miles, because you put too much energy into the first half of your race. And so Amazon is operating on the same thing because if we're fulfilling faster than we said, that means we may be missing promise for other customers because their orders got deprioritized or didn't make it through the pipeline. And I know these sound like extreme cases and kind of like, why would you even worry about this? But when you're operating at the level and the scale of a company like Amazon, this actually matters because there is literally no Slack in the fulfillment supply chain at all. So if something accelerated, that means something else decelerate.
What'd you think? Do you think that's accurate?
Clifford:
For sure. Right. And I mean, the way I think about it for like for a smaller company is you have a cutoff time. And if you're using a three PL or you're using a relatively modern fulfillment center software stack, it's going to do some prioritization for you. But generally speaking, yeah, it's going to take anything that is a priority upgrade in terms of the shipping speed move that in, which could impact promises that you've made two another customer, right. And so if you're not managing that flow and using the promise date, working backwards from what you promised to the customer and you take that as your kind of ultimate, uh, prioritization arbiter, then you tend to make decisions that start kind of like flipping things around. And some customers benefit, some customers don't. But ultimately, again, it's an inefficiency for your, uh, your pick crew.
Clifford:
Now they're on, they have to alter their pick paths. You know, pack out is now like if you don't have fast lanes where you are dealing with those kinds of things in, they're coming down other lanes, you're moving stuff out of the way and trying to deal with that. So yes, without that prioritization and using promise as the work backwards, you really don't have a good opportunity to prioritize. And make the right decisions, both financially or operationally.
Kimberly:
Yeah. Agreed. Agreed. So promise date. So how does Amazon come up with a promise date today?
Clifford:
So how has a pick been picked? Has it been put on a conveyor? Where is it in the building? Has it been slammed? Has it been tendered? They have dozens of pickups per day by dozens of carriers, right?
So they have a level of sophistication that a lot of companies don't have. Yeah. You know, the way you do it is you take the data that you do have and you stitch it together in an end to end way. How many packages we're injected into the fulfillment center today. How many orders were injected into the fulfillment center today? So of those orders, how many were picked, packed and shipped and tendered. And then you get the data from UPS. All this information is available through API or even you can just download an Excel file or something simple and pull that information in so that you can see what the carrier tells you about that tracking ID. So these are all really pretty straightforward and simple things to do. But the most important thing to do is to actually follow a package, don't look at things at an aggregate level because the hides all of the details of what's happening every step along the way in order to get to the package on the doorstep.
Kimberly:
Yeah. Yeah. So you bring up something that's super important as well. And you and I talked about this a little bit previously, so let's talk about carrier performance because that is a big piece of this promise date and it is not something that retailers and eCommerce companies are actively managing. I have had some really tough conversations as an employee and as a consultant where I have gone to the transportation logistics team and said, great, where's your carrier performance metrics? And I get the big blank stare right? And they go, well our accounting team looks at that. They look at that when they do the audit like months from now. So carrier performance has to be a part of the metric if you're going to make promise date, because ultimately that is the person or the company that's going to help you meet that promise, which is, I said, you're going to get it on Tuesday. UPS has to show up on Tuesday to deliver that package. And so my experience is that again, companies stop at ship date because they don't want to get into the complexities of managing their carrier because they don't know what to do if they find out that ups is missing promise and UPS will come back (DHL or FedEx, I'm not just picking on UPS) will come back and say, ‘but we didn't make that promise. You made that promise.’ So let's talk about that a little bit and how you overcome that hurdle.
Clifford:
Yeah, so the other thing that happens a lot is that the carrier data becomes stove piped in the operational group. And there's a certain amount of territorial illness about that information and about the management of the processes of operations that a lot of operational people don't feel comfortable either sharing or putting together into a metrics deck and communicating it and well, let's be honest, there's a lot of executives who just fall asleep when somebody starts talking about this information. So if you wanna like change culture, sometimes it's really difficult because from on high you're really not getting the kind of support you need to like to actually look at that data rocket and then do some changing on it. But I think getting back to just like, what do you do with the data? Once you've got it and you've got information that pinpoints a problem. Every carrier is interested in your business as a seller, and why? So you can turn this information into a kind of thing with your carrier. Talk through, here's what we've seen. You look at the contract contracts typically don't go to to two decimal points of nines and whatnot, but you should be at least looking at the information in your contract, taking the data that you have and then having conversations and then asking for root cause. And oftentimes what you will find are simple things. You have an injection center that's running behind something like that and then ups or your carrier and you can have a really nice conversation about, well, what did that mean for my ability to meet my promise to my customers? A lot of times they will to pay attention to this information, but you have to bring it to them. They're too busy to worry about you unless you actually stand up with the data you have.
And if you do, I've never found anybody from any of the carriers that isn't interested in that information cause it helps them and it helps other customers. So I just think that a lot of companies get complacent about the data or like I'm too small or like they're not going to listen to me and I don't think that's the case at all.
Kimberly:
Yeah, I completely agree with you and I've heard that excuse a lot too. Like we're too small. The biggest excuse I hear out of companies is it's too hard. And you know, when I was working at Nordstrom, this is one of the things I tried to dig into. I was there and the answer I got back was, well, UPS gives us our report two weeks from now. It takes two weeks for us to get a report from, from UPS. And I'm like, Mmm, actually not true. You can get that real time if you have an API hookup and you're getting that data. So I do think that a lot of companies lean back because to your point, they're afraid to go to the carrier and bring up this information and frankly, it's a ton of work. It's a ton of data because just like Amazon, and Amazon measures almost every minute of a package and an FC of an order from the minute it dropped to the minute it was cued to be picked to when it actually got picked to when it got slammed to like when it went into a box. And so each one of those increments is measured down to like the second. And when you talk to a company about measuring a carrier like that, they start to back off. But what you don't understand is that FedEx, DHL, UPS, USBs, they're also measuring their pipeline exactly the same way.
So when you get data back from those companies, a lot of times if you just ask them for the data, you get just this big massive amount of data that you don't necessarily know what to do with. And I think that scares companies as well. But when we worked at Amazon together, I was in worldwide transportation and I sat in a lot of really uncomfortable meetings because I was responsible for international transportation. We're literally the SVPs and the executives would say, why did ups miss? And your initial response is to say, well I don't know. I'm not UPS. Like that's UPS’s problem. No, you have to get on the phone and you have to call them up and find out why did you miss UPS. You cannot wait for UPS to come back and say, Hey, I missed your metric. It's your metric. You have to manage it. You have to drive the carrier to be successful. And I think that is a missing piece. And as you talked about, they're afraid to do it. They don't think that they can do it. They don't think the carrier is going to listen to them. But if you look back at what was going on with carriers 10-15 years ago, and the level at which they perform today, that's driven by Amazon, that is driven by Amazon saying, if you can inject it faster, I'll do injection. If you can't get my ground postal packages to my customer in three days, I'm going to inject into your BMC and cut five days out of your network because you're too inefficient. And that is what I think we're seeing in all of the speed that we see in the carrier networks today.
It was driven by exactly what you said, right? A company calling them out and driving them to do better. And that I think is what the most important piece is for companies to do today. But it's hard, right?
Clifford:
Well it depends on how you define our every order you take as an order ID, if you have any kind of automated fulfillment center software, that order ID is going to translate to a unique package ID. A unique package ID will tie to a tracking. You should be able to chain these three data points from these systems together relatively easily to get, to start to give some information. And even if you start small and you look at it, you eyeball it yourself, like manually just to get started until you can start getting either your BI team's attention or whatnot, but you can get some of this data and put it together and it's really not that challenging.
The other thing I would say that I think is interesting about it is, Mmm, over the course of a year, if you have this information, when your contract negotiations come back with the carrier, this is also really good ammunition for you to be able to pull it all out and say, look this is how we saw your performance. And so it gives you some negotiating leverage if you have clawback clauses in your contracts, it gives you the opportunity to talk about those. And maybe you can recapture some dollars there, but you know, don’t think that you have to go to the level of specificity that Amazon has in order to start getting some data and starting to build those institutional scars around the capturing of that data and the looking at that data and the action of that.
Kimberly:
Yeah. Yeah. So let's talk about that because you bring up a really interesting point because Amazon just didn't wake up one day and say, we're going to start making promise and we're going to promise you're going to get it by Tuesday. Right. They had some data and they started to analyze where the carriers were performing, what the average delivery rates were and started there. And then as they got more and more data, they got more and more precise and they kit and they manage it down to zip code. Okay. So this isn't when Amazon makes you a promise. It is not you're in zone seven. And so we're going to promise you Tuesday. Okay. Amazon looks at the skew, the skew placement, how close is it to you? They look at where this is shipping, right? They look at cutoff. We need to talk a little bit more about cutoff because it's super important, and since they have their zip code, they know what zone you're in. So Amazon is looking at six to 10 elements when they're gonna make you a promise. It is not a flat one to two element piece. And I think that's important for companies to realize. But you can start small. So when we think about how Amazon started first they got data analyzed, it got confidence in what was actually happening and then started to grind away at it. So kind of like what you've been talking about, you can start to get the data, you can start to analyze the data, you can see how your carriers are performing, stitch that together with your FC performance and then start to think about what a promise would look like versus jumping all the way in and trying to do what Amazon does, which is managing a promise date all the way down to an individual skew.
Clifford:
Right. I mean the leap from arrival date to promise date takes a journey of many steps. Yes. You know, to your point it's a marathon, but most companies definitely should be moving from shipped date to arrival date. That's the logical progression. And the only way to get there is to start pulling in the carrier data. If you're not pulling in the carrier data and we talked about not looking at this stuff in aggregate, but if you're not at least pulling that data in, pulling in your FC performance data and then at least looking at that in aggregate as a beginning point, you can't even leap from ship ship date to arrival date.
Kimberly:
Yeah. Yeah. So baby steps, great baby steps. And I agree a lot companies are not out or not managing, they're simply not even managing an arrival date. Right. They're waiting. A lot of companies I see and I work with are waiting for the customer to cue them that something hasn't happened. So one of the things I wanted to bring up that's really, really, really important. And another thing I see lacking out there in the market is companies are not managing data real time. So if they're relying on their customers to give them realtime feedback. They're not relying on their carriers and their data to give them that real time feedback. And that is a big miss in my opinion because you can manage this real time. So when we're talking about looking at arrival date and managing your carrier to see if they've been successful, we're not talking about doing this in a month, three months waiting for your accounting department to come back after they paid the freight bill.
We're talking about getting that data right and looking at it daily because that's when you can make changes that matter. Okay. If we don't do that where if we're looking at it. So you know, one of the things I've seen when I've worked with big luxury retailers and they look at their data two to three months, um, they look at it one, they look at it in aggregate. So now we've, like we've taken some peaks and valleys and we smooth them out because we're like, ah, we had 95% yeah. But yeah, you know what, for two weeks you were at 52% and that, that is the little details that you have to dig into and you have to get real time data to do that. And the other thing that realtime data and looking at that on a daily, weekly, and monthly and quarterly basis is important. Is it? That's what allows you to make the little tweaks, right? That's the thing that allows you to say, this zip code with this carrier isn't that did not perform well last week. We need to change that. Um, and without that, to your point, everything is just peanut butter. And that's where we see companies start to make these big swings. That don't make a lot of sense. Do you see that too?
Clifford:
Oh, for sure. And you know, I have seen and felt that operations teams, even if they have the data or could get the data, are a little reluctant to go nag. There are three PL to nag their carrier and to use the data and every day having a call like this, these things failed yesterday and, you know, keeping them accountable. And to me, that's the essence of what operations is, every day you reset and you get after it again and you try to improve your numbers and improve. And hit all your SLS and whatnot. And if you're looking at it weekly, you're missing all those opportunities. And certainly if you're looking at it monthly, if you're not looking at it at all, shame on you. Yeah. Right. But you know, the job of an operations person is to go nag these individuals and organizations in order to make it better.
We're doing this for the customer. Right. And that’s when you go back to, and we talk about Amazon a lot because we learned so much from them and they're such a great company, but they start from the customer and they work backwards and fulfillment is a manifestation of that. And if you are not every day looking at the information and trying to figure out what went wrong and making sure it doesn't happen again, then you're not really operating in the customer's best interest.
Kimberly:
Yeah. And that was the thing that was at Amazon too. I think they've softened a little, it's not quite as intense as when you and I were there. But if you messed up twice on the same issue, you had a problem that was, we talked about it once. You had an opportunity to fix it. We've talked about it twice. This is on you. If it happened a third time, there was probably not a space for you at Amazon for much longer. But it's that level of like precision. And I just did a podcast on importing like a pro. Basically it's the same thing, data milestones, but the key is action. You have to do something about it. Looking at a metric stack and reporting out that you missed by 0.1%. Yeah, it doesn't matter what, how have you already addressed that issue?
Because you're not helping the customer if you're just looking at data. And I think that's what all of this comes back to is again, starting with the customer, right? Is what you're doing benefiting the customer. And if what you're doing is just looking at a metric stack that is not benefiting your customer. And I do see an operations where I think people get the blinders on and they're worried about their own individual metrics and they stopped thinking about how what they're doing is going to impact the customer. And if you lose that, then I think you, you lose that, um, collective pull in one direction because if everyone isn't pulling towards the customer, then they're pulling for their own needs and that's when things start to break down. So I think you know what, just to summarize, but Cliff and I have been talking about on this particular point is you need real time data.
You have to measure it and you have to look at it on a daily basis and then you have to get on the phone and you have to nag. Which is personally why I think I've been so good at operations because I’m a very good nagger - you have to get on the phone and, and be like, where is it? What happened? Tell me precisely. I want to know the driver who failed. I want the number of the truck that drove off the road. And then the other piece of that too is that there's a feedback loop with this. This is another thing that Amazon used to do around promise date is they, and they still do actually, if they know they're going to miss a promise date, they tell the customer. They don't wait for the customer to call.
Kimberly:
And you and I both been there, a snow storm is closing, a passage truck ran off the road caught fire. Like all of these things happen on a really regular basis. And one you got to try to recover, but two, you got to call the customer or email the customer and say, Hey, we're not gonna make promise because of X, Y, and Z. And that is the responsibility of the retailer, not the carrier. Again, back to starting with the customer. Well, the retailer owns the relationship. Yep. The infamous cold prickly email and Amazon measured that too. How many cold pricklies did we send out? So you don't just get a pass, right? You don't get, uh Oh, we had to change promise because a truck caught fire. You're still responsible for the fact that that didn't get delivered to the customer on time, and now you have a cold prickly, so you got to measure those as well.
Comentarios