Benchmarking Analysis Made Simple for Small Businesses

Learning center series

Benchmarking Analysis Made Simple for Small Businesses

Benchmarking Analysis
Keep reading related articles on Competitor Benchmarking
Start delivering with Metrobi
metrobi-referral
Invite a Business, Get $1000

Most benchmarking advice assumes you have an analyst. It talks about data collection frameworks and peer panels and continuous improvement cycles, and the owner reading it has a van to load at six.

So here is the version that fits a real week. A benchmarking analysis for managing deliveries is six steps, takes one afternoon the first time and about an hour after that, and needs nothing beyond a spreadsheet and your own order records. The output is a short list of specific differences between your delivery promise and your competitors’, with a number attached to each.

This post is the mechanics. For the question of which comparisons earn their place and what to do about the gaps, see how competitive benchmarking helps you stay ahead; if you would rather automate the collection than do it by hand, there is a separate guide to choosing the right competitor benchmarking tool.

The Bottom Line

  • Six steps: pick one question, set your baseline, choose three to six competitors, collect identical fields, do the gap arithmetic, decide and set a review date.
  • Set your own baseline before you look at anyone else. An analysis that starts outward produces opinions about competitors instead of decisions about you.
  • Four numbers are enough to start: on-time rate, cost per delivery, failed-delivery rate, and stops per hour.
  • Express every gap in both units, the measure and the money. “Three points behind on on-time” does not compete for attention. “Three points behind, which is about 40 orders a quarter” does.
  • The first run is for the baseline, the second is for the trend, and the third is the one that changes a decision. Expect that and you will keep going.

Metrobi drivers are rated 4.97/5

"Your delivery drivers actually show up on time and handle products carefully"
— Rachel Parkhurst, Boloco

Trusted by local businesses for:

  • Background-checked professionals
  • Specialized in business deliveries
  • Same drivers for consistency
  • 4.97/5 average delivery rating

What a benchmarking analysis actually produces

A benchmarking analysis is finished when you have a short table with three columns (our number, their number, the difference) and a decision written next to each row.

That framing matters because it sets the size of the job. You are not building a reporting system. You are answering one question, once, with enough rigor that the answer is trustworthy, and then writing down what you are going to do. If the exercise generates a document nobody acts on, the problem is almost always that step one was skipped.

The three types of benchmarking, and which one you need

The textbook split takes thirty seconds and stops you collecting the wrong thing.

Internal benchmarking compares you to your own past. Cheap, immediate, and it is the only kind you can do with no outside data at all. This is step 2 below.

Competitive benchmarking compares you to named rivals. This is the one that explains lost orders, and most of this post is about how to do it.

Functional or process benchmarking compares one of your processes against whoever does it best, in any industry. It is useful occasionally, since a florist can learn route sequencing from a courier, and a poor place to start, because the comparison is interesting rather than decisive.

For managing deliveries, you want internal first, competitive second, and functional only when a specific process is already known to be your weak point.

Step 1: Pick the one question the analysis has to answer

Write it down in a sentence before you open a spreadsheet. Good questions are narrow and tied to something that happened.

“Why did three wholesale accounts leave in the last quarter?” is a good question. “How do we compare to competitors?” is not, because every field is potentially relevant and the analysis never ends.

Other questions that produce usable analyses: Can we justify raising our delivery fee? Should we extend our radius into the next two ZIP codes? Is our order cutoff costing us evening orders? Each one implies a small, specific set of fields to collect, which is the entire point of doing this step.

Step 2: Set your delivery baseline before you look outward

You cannot measure a gap without your own side of it, and most small operations do not have these numbers written down anywhere. They have them as impressions. Four are enough to begin.

On-time rate is deliveries that arrived inside the promised window, divided by total deliveries, for one clean month. If you promise a day rather than a window, measure against the day and note that you are measuring something looser than your competitors may be.

Cost per delivery is every delivery cost for the month (driver time including loading, fuel, vehicle costs, any courier fees, packaging specific to delivery) divided by deliveries completed. Expect this one to be higher than you thought. Nearly everyone underestimates driver time because loading and returning do not feel like delivery.

Failed-delivery rate is attempts that did not complete first time, as a share of attempts. This number is where margin leaks, because every failure costs a second trip.

Stops per hour is completed stops divided by hours on the road. Your route efficiency in one figure, and the one that tells you whether a cost problem is a routing problem.

Two rules make the baseline trustworthy. Use one full month, not a good week. And write down how you calculated each figure, because in a quarter you will not remember whether driver loading time was in the cost, and a baseline you cannot reproduce is not a baseline.

If you want the wider operational context around these figures, last-mile delivery for small businesses covers the ground they sit on.

Step 3: Choose a comparison set of three to six

The right set is the businesses your customers actually choose between when they choose you. Not the biggest names in your industry, and not the national carriers.

Build it from evidence rather than memory. Ask two or three recent customers who else they considered. Check who ranks for your main service plus your city. Ask any account that left where they went. Three to six names is the workable range. Fewer and one unusual operator distorts every row; more and the collection stops fitting in an afternoon.

Include one operator you believe is better than you. An analysis consisting only of competitors you already beat is a morale exercise.

Step 4: Collect the same fields for every competitor

The discipline here is identical fields for everyone, including yourself. A row with gaps in it is a row you cannot compare.

For each competitor, from their own website and a test order: order cutoff time, delivery window offered, service area, delivery fee, order minimum, and whether they show tracking. From their public reviews: the rating, and more importantly the themes in the review text over the last six months, searched for words like “late”, “delivery” and “damaged”.

For yourself, the same fields, plus the four baseline numbers from step 2, which no competitor will have, and that asymmetry is normal. You are comparing promises across the set, and performance only against yourself.

Date every cell. A cutoff time collected in March and compared against one collected in November is not a comparison. The same principle applies when you are comparing storefronts rather than delivery terms, which is what website benchmarking sets out in detail.

Step 5: Do the gap arithmetic

A gap stated in its own units rarely moves anyone. Convert it to money, roughly, and it starts competing for attention.

Take on-time rate. Say your baseline is 92% and a review-text read suggests two competitors are rarely described as late. On 450 deliveries a quarter, 92% means about 36 late deliveries. If your average order is $120, and a quarter of customers who get a late delivery buy less or leave afterward, that is roughly nine accounts at risk, or about $1,000 of quarterly revenue, before the cost of the apology. The arithmetic is deliberately crude. It only has to be right enough to rank the gaps.

Three rules keep the numbers honest.

State your assumptions in the cell next to the figure. “Assumes 1 in 4 late deliveries costs the account” is a claim someone can argue with, which is better than a number nobody can check.

Use ranges, not points, where you are estimating. “Somewhere between $800 and $2,000 a quarter” is more useful than a false precision of $1,340.

Mark which gaps are structural. If a competitor covers twice your radius because they have four vans, the gap is real and closing it is a capital decision, not an operational one. Flag it and move on.

Then rank the rows by money, and look at the top three only.

Step 6: Decide, and set a review date

Every row gets one of four words: close (we are acting on this), ignore (real but too expensive), watch (monitor, do not act), or invert (compete elsewhere instead). No blanks.

Then pick at most two rows to act on this quarter, name who owns each, and put a date in the calendar to re-run the analysis. Two is not a modest target. Moving a cutoff time and tightening a delivery window are both operational changes that touch prep, staffing and routing, and an operation that does both in a quarter has had a busy quarter.

A simple benchmarking analysis template

Six columns is the whole template. It fits on one screen, which is why it gets used.

FieldUsBest competitorGapGap in moneyDecision
Order cutoff4pm9pm5 hoursEvening orders lostClose
Delivery window“Morning”2-hour windowPrecisionComplaints, churn riskClose
On-time rate92%Not described as late~3 points~$1,000/quarterClose
Delivery fee$12$9$3Blocks small ordersWatch
Service area6 ZIPs11 ZIPs5 ZIPsStructural, needs a vanIgnore
Failed deliveries4%Unknownn/a~$900/quarterClose

Copy it into a sheet, add a date in the header, and save the old version each quarter rather than overwriting it. The saved versions are the trend, and the trend is the part that eventually tells you something your instinct did not.

How often to run a benchmarking analysis

Quarterly for the full six steps. Monthly for a two-field check: cutoff time and delivery window across your comparison set, which takes twenty minutes and catches the most common surprise.

Run it off-cycle when something structural changes: a competitor opens or closes, you add a vehicle or a zone, a significant account leaves, or your own on-time rate moves more than two points in a month.

Annual is too slow. Consumers now expect free delivery in an average of 2.7 days, and under one day in grocery and prepared food (AlixPartners). Expectations that move that quickly will outrun a yearly review.

Where small benchmarking analyses go wrong

  • Starting with competitors instead of your baseline. Without your own numbers, every gap is a guess, and the analysis becomes a discussion about rivals.
  • Measuring what is easy instead of what decides. Follower counts and website traffic are easy. Cutoff times and delivery windows decide orders.
  • Comparing against the wrong set. A national carrier’s two-day promise is not a benchmark for a six-van operation. It is a different business model.
  • One run and done. The first analysis has no trend in it, so its conclusions are weaker than they look. Plan three.
  • Too many action items. Fourteen gaps means nothing happens. Two means something does.

Frequently asked questions

How long does a benchmarking analysis take the first time?

An afternoon for the collection and arithmetic, if your order records are in reasonable shape. The baseline in step 2 is the slow part, because it usually means adding up driver hours and vehicle costs for the first time. Subsequent runs take about an hour.

What if I do not have clean delivery data to build a baseline from?

Start collecting it now and run the analysis on a one-month sample in four weeks. Promised window, actual arrival, and whether the first attempt succeeded, logged for every delivery, is enough for three of the four baseline figures. Do not delay the competitive half of the analysis, because competitor promises are public and you can collect those today.

Do I need a benchmarking tool to do this?

No. Below roughly six competitors and a quarterly cadence, a spreadsheet is better, because you can record things no tool can see, like what the account that left told you on the phone.

Start with the baseline

If you do one thing from this post, do step 2. Pick last month, work out your on-time rate, cost per delivery, failed-delivery rate and stops per hour, and write down how you calculated each.

That alone takes an afternoon and changes most owners’ sense of where they stand. The competitive half is faster, and it only means something once your own column is filled in.

About the Author

Picture of Joao Almeida
Joao Almeida
Product Marketer at Metrobi. Experienced in launching products, creating clear messages, and engaging customers. Focused on helping businesses grow by understanding customer needs.
Related posts
In this article
Competitor Benchmarking
Learning center articles
Other Learning Center Subjects