Choosing and working with a 3PL, end to end

A colleague recently asked me what people should actually be looking at with a 3PL, and whether there were any learnings worth passing along. I have written about pieces of this before, but never the whole arc in one place, so this is the full version. The first question is not how to run the process, rather it's whether you should outsource fulfillment at all, and if so, when. The honest answer to when has very little to do with whether you are a big or small brand.

When in the year to do it

A 3PL will tell you onboarding takes one to two weeks. Believe the spirit of that, but put in buffer. The full arc, the eight phases below, plus the issues that surface after you go live, can stretch across months. So time the move for a window where you can absorb a rough patch.

Run your search in your slow season. For most brands that is the first quarter, January through March, once the holiday dust has settled, or the mid-summer lull around July and August after any mid-year spike. Find the gap in your own seasonality. The one rule that holds for almost everyone: do not switch into Q4. Peak is the worst possible time to be ironing out label issues and first-pickup problems. If you are close to it, get through peak on what you know and revisit in the new year.

Phase one: sourcing

Cast a wide net, and only at 3PLs that already serve brands like yours.

Sourcing is harder than it looks, and most people do it too narrowly. They fire off a one-off message, run a Google search, or ask a friend what they use and stop there. The biggest problem I see in this phase is that people do not go wide enough. A referral is a fine place to start that list and a poor place to end it: a name from your network tells you a 3PL exists, not that it fits you, and the most-recommended name is often the best-marketed rather than the best-run. That is its own writeup: why "I know a guy" isn't a 3PL search.

A good 3PL search is at least ten to fifteen real conversations with providers that fit your criteria. Fit is the part people skip. You want 3PLs that already serve customers like you. If you are a DTC brand, they should have other DTC brands. If you are apparel, they should have other apparel. If you are omnichannel, they should already be running omnichannel. A 3PL that has solved your exact problem for someone else is worth far more than one that is willing to figure it out on your account.

To get to that many conversations, use the matchmaking services. Fulfill.com, Third Person, and WareMatch all connect brands with vetted providers, and they mostly cover the US. Other regions have their own matchmakers. You can also lean on people who are connected across a wide range of 3PLs, which is part of what I do. The point of this phase is the wide net, and the discipline to only cast it where the profile already matches.

Phase two: vetting

Ask the questions that get past the proposal, and watch how openly they answer.

This is the phase where it pays most to have someone who has done it before. Every 3PL puts its best foot forward in the proposal. They tell you what you want to hear. Vetting is knowing the questions that get past that, so you can tell whether a provider is actually a fit or just selling themselves as one.

A lot of it comes down to transparency. Ask for references, and watch how openly they answer the direct questions:

1. How many customers do you have, and how many locations?
2. How much square footage are you running?
3. What types of clients do you have, and how many?
4. How many of your clients ship packages with the same weights and dimensions as mine?
5. How many of your clients need the same kind of custom kitting, or whatever the specific work is on my orders?

The answers matter, but so does the willingness to give them. A provider that handles these questions openly is showing you something. One that gets evasive is showing you something too. This only happens on calls, asking real questions and listening to how they respond, which is also why this is the easiest phase to rush and regret.

Those questions test transparency. A second set tests quality, the things that separate a best-in-class 3PL from a cheap rate card and that rarely show up in a proposal:

• Claims recovery. When a carrier loses or damages a package, that is money you are owed, and most brands never collect it. The best partners make filing easy, document claims well, and even tell you when you are under-filing. A weak one treats claims as your problem. Over a year that gap is real money, so ask what their process looks like and what they actually recover.
• Communication. Ask how you will actually talk to them day to day. The strong ones run a dedicated channel with the right people in it, including someone who can walk over and look at your pallet, and they acknowledge a message even when the fix takes a day.
• Capacity to grow into. Can they absorb the version of you two years out without a renegotiation? Automation and real headroom are what make doubling your volume a setup change on their side instead of a crisis on yours.
• Omnichannel. The question is not "do you support Amazon." It is whether adding a channel you have not thought of yet is a connection or a construction project.

What good looks like across all of these is its own writeup: what a best-in-class 3PL looks like, beyond price.

Phase three: analysis

Turn every rate card into your true landed cost before you compare.

This is where a lot of brands get tripped up, because no two 3PLs price the same way and no two rate cards look alike. The most common trap is a provider that quotes very cheap pick-and-pack fees, the front-facing numbers you compare first, and then bakes its margin into shipping. They hand you a clean flat rate card that looks great, but they never convert it into what you will actually pay once surcharges, your real volume, and your package profile are applied.¹

The best 3PLs I have talked to do the opposite. They help you understand how to turn the card into your true landed cost. The hardest part of this phase is comparing providers apples to apples, objectively enough that you actually know who is most competitive rather than who wrote the friendliest-looking card. That is far easier when you are working from source-truth data rather than a summarized export, and it is exactly the kind of normalization we do when we evaluate a setup. If you can, work from your own raw shipping data rather than the version a provider chooses to show you.

Phase four: negotiation

Lead with what matters most to you, and name the offers you have elsewhere.

Negotiation is the combination of knowing what matters most to you and marrying it to what matters most to the 3PL, then finding terms that work for both. Term length is on the table. So is pricing. So are the various specific aspects of the offer. If you have better offers elsewhere, say so, and see whether a given provider is willing to cut its margin a bit to win the business. Leverage you do not mention is leverage you do not have.

Phase five: contract review

Benchmark the terms, and protect your exit above all else.

Everyone is plugging contracts into Claude or ChatGPT now, and that is a genuinely good first pass for surfacing anything concerning. What these tools cannot do well is tell you what is normal, and a contract review is mostly a benchmarking exercise. Without knowing the industry standard, a clean-reading clause can still be a bad deal.

The main thing to watch for is exit. If something goes wrong and you need to leave, you do not want to be on the hook for a large penalty or a minimum you have to keep paying across a long contract. Favorable exit terms are worth more than almost anything else in the document, and a partner confident in their service often defaults to month-to-month rather than a long lock-in with penalties. The other thing worth pushing for is to have the 3PL state its SLAs in the contract itself, so you both have a single source of truth you have agreed on and, ideally, a way to measure it over time.

Phase six: inventory transfer and software setup

Clean your data first, and never move all of your inventory at once.

Once you have signed, you transfer inventory and set up the software, and this is where the work shifts onto you. The brand is responsible for doing this part well, and most of doing it well is data hygiene. Your data needs to be clean, not just in your WMS but in Shopify: HS codes, weights, dimensions, the fields the integration quietly depends on.² The cleaner the data you hand over, the less likely something breaks in the transfer or the setup.

The failure I have lived through is thinking everything looks fine until the inventory physically arrives and then orders will not go out, and now you are disappointing customers for two weeks or going on pre-order. Those have real consequences for a brand. So if you are switching, do not move 100% of your inventory and accept downtime. If you estimate two or three days of downtime, it is very likely to be more. Leave your existing 3PL on good enough terms that they will keep fulfilling for a set period while you transition. Ideally the new 3PL has your inventory sitting in their warehouse for at least a couple of weeks, getting familiar with your setup and running test orders, so they are not scrambling on day one. This is the part of the move where the switching cost the spreadsheet misses actually shows up.

The other half of the transfer is the physical move itself, and every inventory transfer is different. Before anyone schedules a truck, confirm the details on both ends, because the truck you book depends on them: pallet count, average pallet weight and dimensions, whether the inventory is all in one place or sitting in an offsite location, and how the freight gets on and off the truck at each stop. The one that bites people is the loading dock. If a facility does not have a dock, the truck has to be scheduled with a lift gate, and packing it with pallet jacks instead of a forklift changes the time and the equipment you need. A wrong assumption here, like booking a standard truck for a site with no dock, can delay the whole move. So walk both addresses before the pickup date and confirm dock, lift gate, and equipment rather than assuming the last transfer's setup carries over.

The part nobody plans for is the coordination, because a transfer is never one conversation. It is several running at once, between you, the new partner, the freight company, and the outgoing 3PL whose building your inventory is still sitting in. The freight company is usually ready to move the moment it has details, so the bottleneck is almost always the party you are leaving, who is the least motivated to be responsive and often not even on the same tools you are. Get a direct phone number early, decide who owns communication to each party so two people are not chasing the same answer in two threads, and price more than one scenario when the inventory is split across locations rather than assuming a single stop. And keep one eye on the downstream math: what the truck pulls this week determines what is left for every other commitment you have, including any retail or wholesale orders with hard pickup windows. I wrote up the full version of this, the freight booking and the apples-to-apples parcel comparison both, in a quote is only as good as the inputs.

Phase seven: the first orders

Prove it on a few live orders before you turn the volume up.

Then the first real orders go out. The discipline here is the same one that makes carrier onboarding work: prove it in production at low stakes before you scale. Run a test phase on a small slice of live orders, confirm the labels and the data behind them are right, and only then turn the volume up. The week before the first batch is also where you lock in the operational basics, because a package that looks shipped in your system but is not actually moving is the kind of problem you want to find on ten orders, not ten thousand.

Phase eight: SLA tracking and data quality

Choose a partner whose data lets you check both price and performance.

The last phase never really ends. You want a 3PL that gives you genuinely good data quality, because the data is how you check both the pricing and the performance you agreed to. Choose a partner that is tech-forward, shares data openly, and exposes APIs you can pull from to run your own analysis and your own SLA checks. If the data is poor, you have no way to know how your 3PL is actually performing, and that is exactly the position you do not want to be in. This is the same reason to push for source-truth data everywhere else in the stack, and the readout you get from it is only as honest as the delivery data feeding it.

Look in the mirror, and hold onto a good one

Two closing thoughts. First, when something is going wrong, look in the mirror before you look for the door, because the fix is often on your side. Second, when you find a genuinely good 3PL, one that is tech-forward, fair on margin, and willing to work with you, hold onto them. Brands underrate the pain of switching every time. It is a real cost, it always takes longer than expected, and it always introduces risk. A good partner you can build with, who keeps compounding in your favor, is worth more than a marginally cheaper card, which is the whole case for giving your incumbent a real chance before you start this process over. If you want that read on your own setup, that is exactly the kind of call I am happy to take.