How to Evaluate a Logistics Back Office Provider
Pricing models, pilot scope, the questions worth asking — and where these transitions actually go wrong.
Most back office outsourcing decisions are made on a proposal comparison and a reference call, then judged eighteen months later on whether the relationship survived. The gap between those two moments is where the actual decision lives.
This guide covers what to look at before signing: how the pricing models differ and who each one favours, what to ask that a sales conversation will not volunteer, how to structure a pilot that tells you something real, and the failure modes that account for most disappointing engagements.
We provide these services, so treat this as a partial source — but the questions below are the ones we would want asked of us, and a provider who cannot answer them comfortably is telling you something.
One
Decide what you are actually buying.
Three quite different problems get described as "we need to outsource the back office", and they lead to different providers.
Capacity. The work is understood and the process is sound — there is simply more of it than the team can absorb. This is the most straightforward case and the easiest to price. What matters is throughput, consistency and the ability to flex during peaks.
Capability. You need something the team cannot currently do well — freight bill auditing, customs documentation, multi-currency reconciliation. Here domain depth matters far more than headcount, and a cheaper provider without the specific expertise is not cheaper.
Cost. The function works but costs too much. This is the case most likely to disappoint, because cost-driven engagements tend to be scoped tightly, which leaves no room for the exceptions that make freight work messy.
Most real situations are a mix, but one usually dominates. Knowing which one changes what you should weigh in a proposal — and it is worth agreeing internally before you speak to anyone.
Two
The three pricing models, and who each one favours.
Nearly all back office pricing reduces to one of three shapes. Each is legitimate. Each also carries an incentive that is worth understanding before you choose.
Per FTE
You pay for a named number of full-time equivalents. Simple to understand, easy to budget, and the most common model in offshore back office work.
Favours you when volume is steady and you want a team that knows your business over time.
Favours them when volume drops — you are still paying for the seats. It also creates no incentive to make the work more efficient, because efficiency reduces the invoice.
Per transaction
You pay per invoice processed, per document handled, per shipment tracked. Costs scale directly with volume.
Favours you when volume is seasonal or unpredictable, and it aligns incentives on speed.
Favours them on exceptions. A transaction that takes four times as long often bills the same, so the pressure is to process the straightforward and queue the difficult. Ask explicitly how exceptions are handled and priced.
Outcome or gainshare
The provider takes a share of what they recover or save — common in freight audit, where recovered overcharges are measurable.
Favours you because there is no cost without a result, which makes it the lowest-risk way to start.
Favours them over a long horizon — a percentage of recurring recovery can exceed what a fixed fee would have cost. It also biases effort toward what is measurable rather than what matters. Agree how long the share runs and what happens to prevented errors, not just recovered ones.
A practical note: proposals priced on different models cannot be compared on headline number alone. Normalise them against your own expected volume, including a realistic exception rate, before you look at the totals.
Three
Questions worth asking.
Capability questions get answered well by everyone. These are the ones that separate providers, and they are worth asking of us as much as anyone.
- Who exactly will be doing this work, and will they still be here in a year? Back office turnover is high across the industry. A named, stable team that learns your business is worth considerably more than a larger pool. Ask for actual retention figures on comparable accounts, not a company-wide average.
- What happens to an exception at 4pm on a Friday? Freight generates exceptions constantly. The answer reveals whether exceptions are handled or queued — and queued exceptions come back to your team, which is the opposite of what you are buying.
- Who do we call, and what happens if they are unavailable? A single named contact is good until they are on leave. Ask what the escalation path looks like in practice, and whether it reaches someone who knows your account.
- What do you need from us that is not in the proposal? Every engagement requires client-side effort — process documentation, system access, someone to answer questions during onboarding. A provider who says "nothing" has either not thought about it or is not telling you.
- What does leaving look like? Notice period, data return, handover of documented process, and whether anything ends up locked in their systems. Ask before you sign, not when you want out.
- Show me an engagement that did not work. Every provider has them. One who cannot name a single failure is either new or not being straight with you, and the answer tells you how they handle difficulty.
Four
Structure the pilot so it tells you something.
Most providers will run a pilot. Most pilots are scoped to succeed, which makes them poor evidence. A pilot worth running has a few properties:
- Narrow but real. One function, one region or one carrier group — not a curated sample of easy work
- Includes the messy part. If exceptions are excluded from the pilot, the pilot proves nothing about exceptions
- Long enough to survive a handover. Anything under a month tests the launch team, not the ongoing team
- Measured on something agreed in advance. Turnaround, error rate, backlog cleared — decided before you start, not selected afterwards
- Has a defined ending. Both what expansion looks like and what stopping looks like
Five
Where these transitions go wrong.
Engagements rarely fail on capability. They fail on a small number of recurring practical problems, most of which are foreseeable.
The process was never documented
It lived in the head of someone who has done it for nine years. Onboarding surfaces this immediately, and the documentation work lands on your team at exactly the moment they were expecting relief. Budget for it.
The client-side owner left
Engagements depend on one person who knows both sides. When they move on without a handover, the relationship quietly degrades — not through any failure of delivery, but because nobody is left who can answer questions.
Volume spiked during onboarding
Peak season arrived before the team was ramped. This is a scheduling problem, not a provider problem, and it is avoidable by not starting a transition eight weeks before your busiest quarter.
Scope was defined by task, not by outcome
"Process invoices" leaves open who chases the missing POD that blocks the invoice. The gaps between defined tasks are where work falls, and they surface in month three rather than week one.
Nobody agreed what good looks like
Without a measure set at the start, assessment becomes a matter of impressions — and impressions are formed by the exceptions people remember rather than the volume that ran cleanly.
Want to test the questions above on us?
Tell us which function is under most pressure. We will answer every question on this page directly, including where we are not the right fit.
Related