The Year You Cannot Compare To

The years that do not work

A baseline year has to be one where the business, the site, and the measurement were all roughly normal.

Most businesses have at least one year where none of those held.

Using it as a comparison produces a figure that looks precise and describes an artefact.

Which is worth marking in the record rather than rediscovering every time somebody makes a chart.

What makes a year unusable

• A rebuild partway through.
• A tracking gap of weeks or months.
• A change of measurement.
• An exceptional event affecting demand.
• A change to the business itself.
• A domain or address change.

The second is the most common and the least remembered. A tag broken for six weeks produces a year with a hole in it, and nobody who looks at the annual total three years later knows the hole is there.

2020 is the obvious one

And it is worth being explicit about.

For most businesses the pattern of that year bears no relation to any other, in either direction.

Some trades were far busier and some stopped entirely, and both produce a baseline that will mislead for years.

Comparing 2023 against 2020 is not a comparison, and comparing against 2019 introduces a four-year gap in which everything else changed.

The honest approach is to compare against 2022 and note that anything longer crosses an unusable year.

That is a limitation rather than a problem to solve, and stating it is better than working around it silently.

A worked example

A business built a chart comparing every year since 2018 and drew a conclusion about a long-term decline.

Their 2020 was exceptionally high for reasons that no longer applied, which made every subsequent year look like a fall.

They also had four months of 2021 with no tracking after a rebuild, which had been recorded as low traffic rather than as missing data.

Excluding both years, the trend was flat with a modest rise in the last eighteen months.

The apparent decline had been entirely an artefact of two unusable years left in the series.

The chart had been shown to somebody making a spending decision.

Mark them in the record

Which is the whole remedy and takes minutes.

In your spreadsheet, put a note against each affected year saying what happened and whether the figures are comparable.

Not comparable, tracking gap March to June is enough.

Colour or flag the rows if that helps, so anybody building a chart sees it before they build one.

The point is that the person making a comparison in two years will not be the person who remembers, and frequently will not be you.

Every business that has been running a decade has an unmarked hole somewhere in its figures.

Distinguish missing from low

Since the two are recorded identically and mean opposite things.

A month with no tracking shows as zero or as a very low figure, which reads as a terrible month rather than as an absence.

Leave those cells empty rather than entering the recorded number, and note why.

An empty cell is honest and a low number is a lie that anybody averaging will incorporate.

The same applies to a partial month at the start or end of a measurement period.

This is the single most common corruption in a small business's own records.

Use the nearest usable year

Rather than the most recent one.

Where last year is unusable, compare against the year before and say so on the chart.

A two-year comparison stated clearly is more honest than a one-year comparison against a broken baseline.

Where every recent year is compromised, the honest answer is that no comparison is available yet, which people find unsatisfying and is true.

In that case, start the clean series now and be able to answer in two years.

This year will need marking too

Which is the immediate application.

The measurement changes at the end of this month, so 2023 will contain half a year of one basis and half of another.

That makes it an unusable year for any annual total, and a partially usable one if the halves are kept separate.

Mark it now, in the record, before anybody produces an annual figure in January that averages two incompatible things.

Doing it now costs a line and doing it later requires somebody to notice.

Keep the reason with the figure, not in a separate note

A small discipline that decides whether the marking survives.

A caveat written in an email, a meeting note, or somebody's memory does not travel with the spreadsheet, and the spreadsheet is what gets opened in two years.

Put the note in the row itself, in a column beside the number, so that anybody copying the data into a chart carries the warning with it.

The same applies to any figure you export or hand to a supplier, since a clean-looking table invites clean-looking conclusions.

Where the caveat cannot fit in a cell, a single line at the top of the sheet is the next best place and is still better than anywhere outside the file.

The counter-case

Excluding years has its own risk.

A business that marks every awkward year as unusable ends up with no history at all, and some of those years contain real information.

An exceptional year is also worth understanding rather than discarding, since what happened in it occasionally recurs.

And a rough comparison across a compromised year is better than no comparison at all, provided the caveat travels with it wherever it goes.

Mark the unusable years and why, leave missing months empty, compare against the nearest clean year, and flag this one now.

The ten minutes

1. List your years of records.
2. Mark any with a rebuild or gap.
3. Mark 2020 explicitly.
4. Empty the cells for missing months.
5. Note why beside each.
6. Flag 2023 for the measurement change.
7. Compare to the nearest clean year.

Step four is the one that prevents silent corruption, since a missing month recorded as a low number is indistinguishable from a bad month to everybody who looks at it afterwards.

Where the record itself lives is covered in a spreadsheet of your last three years.

Frequently asked questions

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.