Why your rigging spreadsheet will not survive a design review

The problem is almost never arithmetic. In the worked example the spreadsheet's three checks read 22, 50 and 40 percent while the padeye's governing check is 146.5, and the term it has no cell for is larger than the whole allowable.

Updated 7 September 2026 · Companion tool: Lifting Lug Calculator

01

The problem is not that spreadsheets are wrong

They usually are not. Every classical identity in this article is easy to write in a cell and hard to get wrong, and that is exactly why the real failure mode is invisible.

Two arguments against engineering spreadsheets get made often and neither is the interesting one.

"They have errors in them." Some do. But a padeye spreadsheet is three formulas a competent engineer has used a hundred times, and it is usually arithmetically sound. Arguing about defect rates concedes the wrong ground, because the spreadsheet in this article gets every number it computes exactly right.

"They are not verified software." True, and it moves the argument to a place where nobody can win it, because verification is a spectrum and most engineering tools sit somewhere in the middle of it.

The failure that actually gets spreadsheets rejected at design review is different, and it has three parts:

  1. You cannot see what it did not check.
  2. You cannot see where a number came from.
  3. You cannot tell which version produced the result on the page.

All three are properties of the artefact, not of the arithmetic, and none of them is fixed by checking the formulas more carefully.

02

What it did not check

A missing check and a passing check look the same on a printout. In the worked example the missing term is bigger than the allowable it would have been compared against.

A perfectly ordinary padeye spreadsheet: net section, bearing, and weld. A 260 kN lift point on 30 mm plate, sling arriving 30 degrees off the plate axis, with two 10 mm fillet welds 220 mm long and the pin 200 mm above them.

What the spreadsheet computes.

Net section: 47.6 MPa, which is 22.4 percent of the declared allowable.

Bearing: 160.5 MPa, 50.2 percent.

Weld: the force over the throat area, giving 40.4 percent of the 207 MPa allowable.

Three checks, nothing above 50 percent, and a design somebody would sign.

What is not in it.

The sling arrives 30 degrees off the plate's axis, so 130.0 kN of it acts across the plate. That force is applied at the pin, 200 mm above the weld, so it puts a moment into the weld group that the spreadsheet has no cell for.

Against an allowable of 207 MPa. The term with no cell is larger, on its own, than the whole allowable it would have been compared against.

Lifting Lug Calculator · computed at page render

The same padeye, with every check performed

Identical geometry, identical load, identical declared allowables.

Net-section tensionthe spreadsheet said 22.4%22.4%
Pin bearing on the lugthe spreadsheet said 50.2%50.2%
Double-plane shear-outnot in the spreadsheet69.1%
Pin double shearnot in the spreadsheet25.8%
Fillet weld throat resultantthe spreadsheet said 40.4%146.5%
Fillet weld von Misesnot in the spreadsheet120.3%
Governing check: Fillet weld throat resultant (mechanics)146.5% utilisationFail

The two checks the spreadsheet did share reproduce exactly. Its arithmetic was correct throughout, and its answer was wrong by a factor of three and a half on the check that governs, because that check was not in it.

Open this example in the calculator
The three bars a spreadsheet printout would carry, and the six a full check produces. The tallest bar is not one of the three.

A missing check has no visual signature. That is the whole of it. A spreadsheet that omits shear-out prints exactly the same as one that includes it and passes, and a reviewer who does not already know shear-out exists has no way to notice.

03

Where the numbers came from

A review is mostly an audit of provenance, and a spreadsheet is the artefact least able to supply one.

Take the same calculation and ask the questions a reviewer actually asks.

Where did the 260 kN come from? A cell. Typed by somebody. Was it weighed, was a contingency applied, was a dynamic factor applied, and if so was it applied here or upstream? The cell does not say, and the two most common load-path errors - a factor applied twice, or one everybody assumed somebody else applied - are both invisible in a number.

Where did the allowable come from? Another cell. A design route? A material standard? A previous project? A memory?

Which edition? Design standards change. A spreadsheet built against one edition looks identical after the edition rolls.

Which version of the spreadsheet? The one on the network drive, the one in the email, or the one with the fix. All three exist and all three print the same.

What was out of scope? The genuinely interesting question, and the one a spreadsheet has no mechanism for at all. A calculation that cannot say what it did not attempt cannot be reviewed for what it missed.

A calculation record has to answer all five without the author in the room. That is not a high bar for software and it is a very high bar for a spreadsheet, because it requires discipline that nothing enforces.

What answering all five looks like. The same padeye, as a record rather than as cells. Every question above is answered somewhere in the document below, which is the only claim being made for it:

padeye-review-example-rev-A.pdf

Lifting Lug Calculator · complete, unwatermarked

Open these inputs
Load and its factors at the top, geometry with the drawing it describes, the allowables and where they came from, every check with the source it traces to, then the warnings, the assumptions and what stayed out of scope. Scroll it and try the five questions on it.

04

When a spreadsheet is exactly the right tool

More often than this article's tone suggests. The failure is not using one; it is issuing one.

Working out the size of an answer. In a meeting, on site, deciding whether an idea is worth pursuing. Fast, transparent, and nobody is going to build from it.

A calculation you have derived and can show in full. If the whole derivation fits on one printed page and every input is on that page, a spreadsheet is a fine vehicle for it.

Parametric exploration. Twenty variants of one arrangement to see which lever matters. This is a spreadsheet's strongest use and it is genuinely hard to do any other way.

A calculation with a documented, tested basis and a controlled release. Spreadsheets can be verified, version-controlled, protected and reviewed. Organisations that do this well have a procedure, an owner and a test set for each one. Organisations that say they do this and have a folder called "calcs" do not.

A spreadsheet is a working tool. The failure is issuing the working tool as the deliverable, because a deliverable has to carry its own provenance, its own scope statement and its own version for a reader who cannot ask the author anything.

05

Making a spreadsheet survive review

Nine habits. They are not about the formulas.

If the spreadsheet is going to be issued anyway

  1. 01Put every input on the first page with its sourceValue, unit, and where it came from. A weighbridge ticket, a drawing revision, a standard clause.
  2. 02State the design route and its editionNamed on the sheet, so a reader can tell which document's constants are in it.
  3. 03List the checks performed, as a listNot as a series of rows that happen to exist. A reader should be able to count them.
  4. 04List what was not checked, and whyThe single most valuable line on a calculation, and the one a spreadsheet never has.
  5. 05Show the formula, not just the resultEither printed alongside or written out in a derivation sheet. A cell reference is not a derivation.
  6. 06Version it visiblyA version number and date in the footer of every printed page, changed when the file is.
  7. 07Name the author and the checkerTwo people, on the sheet, and the checker did not build it.
  8. 08Test it against a known caseOne worked example with a published or independently derived answer, kept in the file and re-run after every change.
  9. 09Lock the formulasSo that the copy that arrives at review is arithmetically the copy that was tested.

06

Seven ways a spreadsheet fails a review

None of them is an arithmetic error.

1. A check is missing and nothing shows it. The worked example: three correct numbers and a governing check at 146.5 percent that was never computed.

2. An input has no source. The reviewer can see the number and not what it is.

3. A factor is applied twice, or not at all. Invisible unless the load path is written out with each factor's owner named.

4. The allowable has no basis. A number in a cell, possibly from a previous project, possibly from a superseded edition.

5. The version is unknown. Three copies exist, they print identically, and one has a fix in it.

6. The scope is unstated. Nothing on the sheet records what was deliberately left out, so a reviewer cannot tell an omission from a decision.

7. The author is the only person who can read it. Which makes the review a conversation rather than an audit, and conversations are not records.

Common questions

What is actually wrong with engineering spreadsheets?
Usually not the arithmetic. The three things that get them rejected at review are properties of the artefact rather than of the formulas: you cannot see what it did not check, because a missing check and a passing check print identically; you cannot see where a number came from, because a cell holds a value and not a provenance; and you cannot tell which version produced the result, because three copies of the same file print the same.
How much can a missing check matter?
In this article's worked padeye, a spreadsheet with net section, bearing and weld shear reports 22, 50 and 40 percent, and the same lug's governing check is 146.5 percent. The missing term is the moment the off-axis sling force puts into the weld group through the lever arm from the pin, and on its own that term is 228 MPa against a 207 MPa allowable. The spreadsheet's arithmetic was correct for every quantity it computed.
When is a spreadsheet the right tool for a rigging calculation?
For working out the size of an answer in a meeting, for a calculation you derived and can show in full on one page, and for parametric exploration of twenty variants to find which lever matters - which is a spreadsheet's genuinely strongest use. It is also fine as a controlled, tested, version-managed tool where an organisation actually has that procedure. The failure is not using one; it is issuing one as the deliverable.
What does a reviewer look for that a spreadsheet cannot show?
Five things: where each input came from, where the allowable came from, which edition of the design standard the constants belong to, which version of the file produced the printout, and what was deliberately not checked. The last is the most valuable line on any calculation and the one a spreadsheet has no mechanism for, because a calculation that cannot state its own scope cannot be reviewed for what it missed.
How do I make a rigging spreadsheet reviewable?
Put every input on the first page with its source. Name the design route and its edition. List the checks performed as a countable list, and list what was not checked and why. Show the formulas rather than only the results. Version it visibly in the footer of every printed page. Name an author and a separate checker. Keep one worked test case with an independently derived answer in the file and re-run it after every change. Lock the formulas so the copy at review is the copy that was tested.

Sources

Every document below is linked at its publisher or regulator. Xarpis reproduces no standard text; where a clause is named, the identifier is given so you can find it in your own copy.

  • LOLER 1998Lifting Operations and Lifting Equipment Regulations

    UK Health and Safety Executive · free to read

    The UK duty framework for lifting operations: planning by a competent person, supervision, and thorough examination of lifting equipment and accessories. Like OSHA's rules it governs the process, not the arithmetic.

  • ASME BTH-1Design of Below-the-Hook Lifting Devices

    ASME · paid document

    Structural, mechanical and electrical design criteria for below-the-hook lifting devices, used alongside ASME B30.20 which carries the safety requirements. The current edition is BTH-1-2023; Xarpis implements the 2020 edition and says so on every result.

  • EN EurocodesEurocodes: Building the future

    European Commission, Joint Research Centre · free portal

    The Commission's own Eurocodes portal: the structure of EN 1990 to EN 1999, the database of Nationally Determined Parameters, and the second-generation timetable. The standards themselves are sold by the national bodies, but the NDP database is free and is what decides which partial factors apply in your country.

  • ANSI/AISC 360Specification for Structural Steel Buildings

    American Institute of Steel Construction · free to read

    The US steel design specification, in both LRFD and ASD. AISC publishes it for free download, which makes it one of the few structural standards a reader can check the same afternoon they read about it.

Run the check properly

Reading about a calculation is not the same as being able to hand one over. These tools produce the traceable record.

Something here wrong, or thinner than it should be? Tell us which paragraph and it gets rewritten. Articles carry the date they were last revised for exactly this reason.

Engineering spreadsheet errors a review will find · Xarpis