Sorting & Grading - First Steps in Refurbishment

What really happens after a product is returned? It all starts with sorting and grading—the core operational steps that determine everything else in the refurbishment flow.

In this clip from the ReCommerce Podcast, Tuomo Laine (CEO of TWICE Commerce) explains how businesses evaluate returned items, decide what to do next, and calculate whether it’s worth investing in refurbishment—or not.

Covered in this clip:

  • Sorting: How companies decide which process line an item goes to based on product type, brand, or visible condition
  • Grading: Professionally evaluating condition to determine resale potential, repair needs, and cost
  • How grading logic evolves through real process data and market demand shifts
  • Why this phase determines your margin—and why it must be tracked on an item level
  • The role of timestamped records, employee time, and spare part costs in accurate grading decisions

Karri: Alright, so the first phase was to get the license plate number on the item, or ideally it's already in the item and maybe it's already put into the system. And then the first step, I guess, is this type of general sortation or similar.

Tuomo: Yeah, so in that sortation, we try to do those high-level decisions on, for example, which process line this item goes. So in many cases, if you are a professional refurbisher, you might have established relationships with brands where you have requirements like that for, I don't know, Levi's, you need to do certain kind of repairs; for newly changed, you do something else. So you might have staff that are specifically trained for specific product brands or models or electronics. It might be like, you handle tablets, you handle phones, you handle computers, because there's some level of specialized knowledge that is needed there. So in sortation, we're trying to figure out on a high level which kind of process line this item would go. If you're super specified and you know that I only receive this kind of products, the sortation becomes maybe a bit easier. So that's a big thing. And then it might be something related to the condition of the product. So I might already see that this, let's say, the laptop that came in is in 100 pieces, it's broken, it's like there's no way that our process line will be able to do anything related with it. So I might just disregard it and recycle it. So you can do this decision without even knowing too much about the brand. Similarly, if it's like, let's say, an outdoor tent, if it's extremely dirty, you know it was used in Glastonbury, comes back, you look at it and you know that, okay, so before anyone can do anything with it, it needs to go through a wash. So you might do this decision already at sortation. So it's these high-level decisions to figure out where stuff goes and whether there needs to be a preface before going there, for example, like washing.

Karri: Alright, so now we have the license plate and the initial sortation that can be quite unique already, but the main idea is to have like this high level, which type of line is it going if they are very separate, or if it's going already to the next use case, or if it needs some kind of washing or something preface like that. And after that, if, let's say that in this case it's a good item, we want to proceed with the refurbishment. What is the next step?

Tuomo: Sure, and we probably need to revisit the next best use because there's also a little bit of these decisions there. After we get to the kind of right process line, so again what that means is that it's now at a process line in a condition that someone can start working with it. The item likely has some level of declaration, meaning what is it and some initial condition. For example, with phone refurbishment, it might have consumer-attached declaration to it, like, "I've said that this is in good condition." So this grading process might start to work against that. So there's a professional grading happening there independently and then they might compare and that might affect your cashback or trading value or so. But if we put that aside for a while and we assume that all of this is happening without that kind of a value proposition, now I have the item in front of me and my job is to start figuring out what grade is it. And grading guides come from manufacturers, brands, retailers, or if you're kind of operating this internally, you've defined few grading categories. One of the reasons why you have this is to figure out is it worthwhile for me to invest how much to this process. Like you might have an estimate that is continuously improved through actual data from the process and actual sales, that how much does it cost, for example, for me to take a Grade 4 item and to refurbish it to be Grade 1. That, of course, has some cost to it. I might have to use some spare parts. But then of course, you have a margin count into it that from a Grade 1 phone you can maybe or item you can make a bit more money. So it's this kind of a condition grading process that starts from there. Your overall goal, of course, is to do this analysis that how much what is needed in order to improve grade, the repair cost of that, and that calculates everything from like cost of washing, cost of spare parts, cost of employee time. And then doing this analysis that to what level should I grade it? Do I just change the screen? Do I start to put a little bit more memory there? Do I kind of start to remanufacture it from the beginning? In most cases, if you're trying to scale your operation, I would say that this is the place where margins are made or lost. And the analysis that you can do there, like how good on granular can you go there. And the things that I mentioned is that like the cost basis are not only the spare parts, it's also the employee time. So this is now starts to be the reason why every item and the overall process and the timestamps kind of that happens that need to be followed on item level. Because first you start with an estimate and when you're going to day zero, you're just saying, okay, like I suspect that it might take 20 minutes and these spare parts to push a Grade 4 to be Grade 1. As you start to actually do it, you track your operational efficiency and then you start to see what the real costs are and then you kind of adjust your grading logic, which might then start affecting the decisions on hey, what should happen to this item now. So that's I think the most interesting part and that's probably the one where you can make most value based on your knowledge of the item. Like your supply of spare parts and all of these things. That's where most value is created.

Karri: And I guess one additional variable is also the, let's say if you're, for example, reselling the item, is also the demand for it like on different grades and that's something that can actually change quite a bit. So we've been talking a lot about iPhones and for example when there is a new model the previous models might kind of like start to lose value on...

Tuomo: That's just value goes down.

Karri: Yeah, yeah, so that's something that also needs to be taken into account that this same phone if I actually refurbish it to be Grade 1, a month ago could have been profitable, but nowadays it's not and now it's maybe better to do it to Grade 2 or 3 and then sell it a little bit cheaper.

Tuomo: Definitely, yeah, so this really goes into the category of whether it's in repairs in budget or no. I think before going to the idea that okay so we now refurbished it and then we're going to put it out to be sold, I think it's worthwhile to figure out what happens if our analysis says that it's not worthwhile. Not worthwhile to grade it at all. So either it was, and this might have happened already at the sortation, that we defined that this is effectively junk. Or, and it has only residue value that I'm going to talk about. And then it might happen here in the condition grading that a more trained process actually defines that, okay, there is refurbishment potential here, but as we count in the cost of refurbishment, then we actually end up saying that the economics don't make sense. Physically we could do it, but economics do not make sense. So from both of these paths, the item might go to next best use. So figuring out what do we do with it. So again, super specific. It might be that let's send it back to the customer saying, "Hey, do you want it back or do you want us to recycle it?" Now that's when the analysis starts to happen that okay, so even though it might not economically make sense to push this item to be resellable or reusable, some parts of the kind of spare parts that is attached to it might still make sense. So I might take the battery out, maybe there's value there, I can use it somewhere else. Maybe the screen is still in good condition, everything else is just broken due to water damage or whatever. Or with clothes, you know, there's maybe some buttons here that we can use in other. So I can start to figure out I might have a specialized process for it, or it might happen in the actual process line, but you can start kind of then figuring out next best use like ripping out all of these spare parts that you might use later. Again, aggregately then improving your margins, because then you don't have to buy those spare parts, you can use it from those you kind of got it for free. Or it might be material recycling. So if you are vertically integrated as a retailer, maybe it just goes back to your manufacturing where they, you know, turn materials back to fiber and all of that, like material flow optimization. Or it might be that it goes to Goodwill, just donating it and all of these things. But it's this analysis and super specific again per product category, per operator, whether this is an aggregated refurbisher, whether it's in-house. This analysis depends a lot on that.

Karri: Okay, so the next best use, you can kind of like at any point that you are doing the check, so it can be the initial check, it can be the grading, or at any time it can kind of like go into the same bucket.

Tuomo: Yeah, and you're just doing it like on a different level. Like initial sortation you're seeing that it's like bent and cracked and all that, you might be immediately like, okay, yeah, we need to, we're not able to repair it. And then in the later phase, if there is some, let's say, like engineer actually looking inside a phone or something and then they can say, okay, there is a bunch of water damage and now it goes to the next best use. Yeah, and it's, it's good to components. There's the physical analysis and the economical analysis. So sometimes the physical analysis makes sense to start refurbishing it, but the economics do not make sense. Like you said that the model or such that the end market value has just plummeted, but there are super interesting categories. Like there's a Finnish company called Kameras that refurbishes old film cameras. And almost the film cameras are so rare, it's like collectibles. So yeah, the market might not right now be there, but you might have to foresight and say that this is still a quite real model, so we will hold on to it as a... Not only as spare parts, but because we're gonna put it on the shelf because maybe the kind of value's going up as it's becoming retro or vintage or similar. So with this...

Karri: And I guess that's also a good example that the example refurbishment flow that we have, we have here is kind of like very simplified version. So you actually have all the spare parts and you might have this storage phase and that's probably very unique to all, all companies.

Tuomo: Definitely. And this, I think this is, this, even though it's a small part there in the, in the, in the graph that we have, but this condition grading analysis is probably the one where software can do a lot of heavy lifting for you. It's probably hard to have one software that does it all like market analysis and all that, but I would say that the requirements for your software should be such that there's quite a lot of storage, you can store a lot of data per article level, on a time stamp level, in order you to do process analysis and in order to get entire that data to other software that are more specialized, for example, pricing optimization and so on. So it is probably a complex commerce in general. I think is always a little bit of a spacket of products. You kind of have to, it's rare that you can run everything in one. You kind of usually have some add-ons or more apps that you use and all of that. So I think this is the same thing. You need a system that is highly inter-integratable to other software to do an optimization on that grading phase.

Karri: Okay, but all the inventory system, I guess the main point is that you are able to track every item individually, you're able to have a timestamp on different phases, when it enters here, when it leaves here, and then some ability to input like expenses there, like maintenance cost or spare parts or similar.

Tuomo: Yeah, definitely. So it's kind of, yeah, all of that. And maybe the one thing that I would add is that the system that you have needs to acknowledge that you get items back, kind of, that there's, that the moving between phases is not necessarily always a one-way street. But you might in practice have items go like through bouncing around something like a flowchart that we have here from different phases to another. And if your system assumes that it's kind of a linear progression of like, okay, enter, do something, exit, it rarely is that smooth. So there's probably some circulation between phases there that happens. So you want a system that doesn't force you to run your refurbishment process in one way because you're likely continuously adjusting that to improve your operational efficiency and margin. So the flexibility needs to be built into the software.