> ## Documentation Index
> Fetch the complete documentation index at: https://help.owlery.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Why Owlery Uses the Midpoint

> Why spot savings is benchmarked against the median of all bids rather than the average or second-cheapest.

**Where to find it:** `Analytics → Metrics Overview → Operations Tab`

Spot savings compares the rate you tendered against the **midpoint** — the median — of every bid you received on that quote. For how to read the dashboard, see [Spot Savings](/shippers/spot-savings). This page covers why the midpoint is the benchmark.

## The Question the Metric Answers

Spot savings answers one question: **what would this load have cost without a competitive bidding process?**

Without one, you'd have called a broker or two and taken what came back. It's fair to object that you'd still have taken the cheaper of those two calls, which would put you below the midpoint of five bids rather than at it. But that objection assumes those brokers would have quoted you the same numbers either way, and they wouldn't have.

A carrier bidding into an open event knows it is being compared. A carrier quoting a shipper who rang them directly knows it isn't. Running the event doesn't just show you more of the market — it changes the prices the market shows you. So the right comparison isn't the cheapest quote from a smaller sample. It's an ordinary price from a market that was never under pressure to sharpen its pencil, and the midpoint is Owlery's estimate of that price.

## The Benchmarks Considered

Four benchmarks were considered. Applied to the five-bid quote in [A Worked Example](/shippers/spot-savings#a-worked-example) — Carriers A through E bidding \$950, \$955, \$1,100, \$1,200, and \$2,400, with Carrier A's \$950 tendered and \$1,100 the midpoint:

| Benchmark                          | Savings it would report | How it holds up                                                                                                                           |
| ---------------------------------- | ----------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
| Second-cheapest bid (\$955)        | \$5                     | Answers a different question, and behaves backwards as a headline number — see [On the Second-Cheapest Bid](#on-the-second-cheapest-bid). |
| Most expensive bid (\$2,400)       | \$1,450                 | Assumes you'd otherwise have picked the worst rate on the board, which nobody would defend.                                               |
| Average of all bids (\$1,321)      | \$371                   | A single runaway bid drags it. Drop the \$2,400 quote and reported savings fall substantially.                                            |
| **Midpoint of all bids (\$1,100)** | **\$150**               | Assumes neither the best nor the worst case, and an outlier bid moves it far less than it moves the average.                              |

## On the Second-Cheapest Bid

Measuring against the second-cheapest bid is a reasonable answer to a *different* question: what would this load have cost if my best carrier hadn't shown up? That's worth knowing, and it's a fair way to size how exposed a lane is to a single carrier. It just isn't a measure of what competitive bidding is worth to you.

It also behaves backwards as a headline number. In that example, Carrier A and Carrier B came in at \$950 and \$955, while the rest of the field sat at \$1,100 and above. Against the second-cheapest bid, that load reports \$5 saved. Against the midpoint, it reports \$150. The \$5 figure is small precisely because you got two of your carriers to fight hard, which means the measure shows the least impact when your event worked best. It also only ever looks at the top two bids, when what you built was competitive pressure across the whole field. That field is what is worth measuring against.

**It also depends too much on who happened to show up.**

Suppose Carrier A sits that event out — truck already booked, dispatcher missed the email, nothing to do with your lane. The other four bids come in unchanged, and you tender Carrier B at \$955.

| On this load            | Carrier A bids | Carrier A sits out |
| ----------------------- | -------------- | ------------------ |
| What you paid           | \$950          | \$955              |
| Second-cheapest reports | \$5            | \$145              |
| Midpoint reports        | \$150          | \$195              |

You paid five dollars more, and the second-cheapest measure rewards you with twenty-nine times the savings. Nothing about the lane changed; one carrier didn't answer the phone.

That's the deeper problem. Second-cheapest rests on a single bid, and *which* bid that is depends entirely on who turned up. Lose one carrier and the benchmark relocates. The midpoint rests on the shape of the whole field, so the same missing carrier nudges it instead of moving it. It does rise — from \$1,100 to \$1,150, lifting reported savings by about 30% — because with the cheap end of the field gone, the market you'd have faced without an event genuinely was more expensive. That's a real shift, sized to a real change. The 29x jump is not.

This is the ordinary reason to reach for a median instead of a single ranked bid: medians are hard to push around. A benchmark should move when your market moves, and sit still when it doesn't.

## Why This Number Is Almost Always Positive

Worth saying plainly: if two or more carriers bid and you tender the cheapest one, you are below the midpoint by construction, so spot savings will be positive nearly every time.

That is the expected result, not a thumb on the scale. Running a competitive event almost always beats not running one, and the metric is sizing *how much* that was worth on a given lane rather than testing *whether* it worked at all. The number that can genuinely go against you is [**Avg Bids per Spot Load**](/shippers/spot-savings#the-summary-tiles) — read the two together.