Two-Way Matching Looks Simple. Here’s Why It Eats So Much Time

Two way matching sounds simple. Until you actually do it.

If you’ve ever explained two way matching to a client, you probably start confidently.
“Invoice matches the purchase order. Done.”

Then you pause.
Because you know that’s not how it plays out in real life.

Two way matching looks neat in a textbook. In practice, it’s one of the most time-hungry, mentally draining parts of accounts payable. Especially for bookkeepers and small businesses dealing with physical goods, freight, partial deliveries, price changes, or messy supplier behaviour.

Let’s talk honestly about the time and difficulty involved. Not the theory. The real day-to-day grind.

A quick reset. What two way matching actually means

Two way matching compares:

• The supplier invoice

• The purchase order

You’re checking that quantities, prices, tax, and totals line up before approving the bill.

No goods receipt involved. That would be three way matching. This is “simpler.” On paper.

But even two documents can create a surprising amount of friction once you factor in how suppliers actually behave.

The hidden steps nobody counts

Ask a junior accounts clerk how long two way matching takes and you’ll hear something like “a minute or two.”

Ask someone who actually does it all day, every day, and you’ll get a laugh.

Here’s what really happens, step by step.

You open the invoice.
You open the purchase order.
You realise the supplier used their own item descriptions, not yours.
You scan line by line to map things mentally.
You notice the quantities are close but not exact.
You remember there was a partial shipment last week.
You search emails.
You double-check freight was meant to be billed separately.
You wonder if GST applies to every line.
You hesitate.
You either escalate, guess, or park it for later.

That’s not two steps. That’s a chain reaction.

And it happens dozens, sometimes hundreds of times a week.

Time adds up faster than you think

Let’s be conservative.

• Simple, clean match: 1 to 2 minutes

• Slight discrepancy: 5 minutes

• Partial delivery or price change: 10 minutes

• Freight invoice tied to multiple POs: 15 to 20 minutes

Now multiply that.

A business processing 300 invoices a month with even light two way matching complexity can easily burn:

• 15 to 25 hours per month on matching alone

That’s half a work week. Gone.
Not filing. Not advising clients. Just reconciling reality with paperwork.

Bookkeepers feel this pain the most. You’re the one cleaning it up, often without context, weeks after the purchase happened.

Where two way matching really breaks down

The difficulty isn’t matching.
It’s exceptions.

And almost every real invoice is an exception in some way.

Partial deliveries

Suppliers ship what they have, invoice what they ship, and reference the full PO anyway.

Now you’re doing mental gymnastics:

• Was this the first shipment?

• Is another invoice coming?

• Should I reduce the PO balance?

• Do I wait or process now?

None of that is written on the invoice.

Price changes no one told you about

Costs move. Fuel surcharges appear. Supplier pricing quietly shifts.

The invoice is “wrong” compared to the PO, but it’s also correct in reality.

So what do you do?

• Reject it?

• Adjust the PO?

• Ask the client?

• Code it and move on?

Every option costs time.

Freight invoices are their own special headache

Freight almost never matches cleanly.

One invoice might relate to:

• Multiple purchase orders

• Multiple suppliers

• Multiple tax treatments

Trying to two way match freight using basic invoice capture tools is where most systems fall over.

This is where tools like Dext and Hubdoc start to show their limits. They capture the invoice well enough. But matching logic is still mostly manual.

You end up splitting lines, cross-referencing spreadsheets, and hoping nothing gets missed.

The mental cost nobody talks about

Two way matching isn’t just slow. It’s cognitively heavy.

You’re constantly switching context:

• Reading documents

• Remembering past conversations

• Interpreting intent

• Making judgement calls

That kind of work drains energy fast.

By mid-afternoon, accuracy drops. Not because you don’t care. Because your brain is tired of exceptions.

This is where errors creep in:

• Overpayments

• Duplicate processing

• Incorrect coding

• Missed accruals

And those errors come back later, usually during BAS or year-end, when nobody has the patience to revisit them.

Why small businesses struggle even more

Large enterprises design processes around matching.
Small businesses inherit chaos.

Purchase orders might be:

• Raised after the fact

• Raised for round numbers

• Raised by someone who no longer works there

Invoices arrive anyway. And accounts payable has to make it work.

For small teams, two way matching becomes a bottleneck because:

• There’s no dedicated AP role

• Knowledge lives in people’s heads

• Approvals happen verbally or on Slack

Bookkeepers then step in, often without access to the full story.

You’re matching documents, but also reconstructing decisions.

Why automation hasn’t fully solved this yet

Plenty of tools claim to automate two way matching.

Most do one of two things:

• Match totals only

• Flag differences without context

That still leaves you with the hard part. Deciding what’s acceptable and what’s not.

True two way matching automation needs to:

• Match at line level, not just totals

• Understand tolerances

• Recognise common supplier patterns

• Handle multi-account and multi-tax scenarios

• Learn from past approvals

This is where bookkeeping AI and accounting AI start to matter. Not OCR. Not data capture. Actual decision support.

What good automation changes day to day

When two way matching is handled properly by software, a few subtle but powerful things happen.

You stop opening two documents side by side.
You stop re-checking the same supplier every month.
You stop second-guessing obvious approvals.

Instead:

• Clean matches pass through quietly

• Edge cases are surfaced with context

• You review, not reconstruct

The time savings are real, but the mental relief is bigger.

Bookkeepers often say the biggest win isn’t speed. It’s confidence. Knowing that what’s in the ledger reflects what actually happened.

A realistic view on time saved

Let’s ground this.

Good two way matching automation won’t eliminate work. It reshapes it.

What you can expect:

• 70 to 80 percent of invoices auto-matched

• Review time cut from minutes to seconds

• Exceptions grouped logically instead of scattered

For a 300-invoice month, that can mean:

• 10 to 15 hours saved

• Fewer client follow-ups

• Less end-of-month stress

That’s time you can use for higher value work. Or just finishing on time.

Why this matters more than ever

Supplier invoices are getting more complex, not less.

More freight.
More split billing.
More overseas suppliers.
More tax nuance.

At the same time, bookkeepers are under pressure to:

• Handle more clients

• Maintain margins

• Deliver faster turnarounds

Manual two way matching doesn’t scale with that reality.

And pretending it’s “just part of the job” is how burnout creeps in.

The quiet shift happening now

More firms are rethinking how much human energy they spend on matching.

Not because they want to cut corners.
Because they want consistency.

Accounting AI isn’t about replacing judgement. It’s about removing repetition so judgement can be applied where it actually matters.

Two way matching is a perfect example.
It’s rule-heavy, pattern-driven, and exception-focused.

Exactly the kind of work software should carry most of the load for.

Final thought

Two way matching isn’t hard because it’s complex.
It’s hard because reality doesn’t behave.

Every mismatch tells a small story about how businesses actually operate. The trick is not forcing those stories into rigid systems, but building systems that expect them.

If you’re feeling like matching takes longer than it “should,” you’re not doing it wrong.

You’re just seeing the full picture.

And honestly? That’s the part worth fixing.