Tags, Projects and the Difference
Two structures that look similar and behave differently over years. Using them for the wrong thing is what makes old records hard to query.
Most tools offer both a category and free tags. The distinction between them is rarely explained and it determines what your record can answer in five years.
To try the approach with real entries rather than memory, start with the full article. The result should remain easy to correct, export and interpret after the immediate details have faded.
What each is for
A category is exclusive and stable: one per entry, chosen from a short fixed list, meaning the kind of activity.
The broader practice of time and billing records is also covered by Harvest, which is useful when designing a record that should remain understandable later.
A tag is additive and volatile: any number per entry, freely created, meaning anything you want to be able to find later.
A project is a specific instance with a beginning and an end, which behaves like a tag rather than like a category.
Confusing the three is the commonest structural mistake in long records.
The failure it produces
Projects used as categories: forty dead labels, no stable axis, nothing comparable across years.
Tags used as categories: entries with three tags and no way to total anything without double counting.
Categories used as tags: a fixed list that grows to sixty entries and stops being consistently applied.
The arrangement that works
One category per entry, from five to eight stable options.
Any number of tags, created freely, never used for totals.
Project as its own field if you need it, or as a tag if you do not.
The category carries the long comparison. Everything else carries the detail.
Tag hygiene over years
Tags proliferate. After three years you will have hundreds, many of them near-duplicates: "client-call", "clientcall", "calls".
Which is tolerable, because tags are for finding rather than for counting.
But an annual tidy — merging obvious duplicates, in the analysis rather than in the data — makes them usable.
Keep a list of the tags you actually use, or you will invent a new variant every time.
What tags are genuinely good for
Marking anomalies: "ill", "travelling", "outage". These explain the shape of a period years later.
Marking experiments: a period when you changed something and want to find it again.
Cross-cutting concerns that do not fit the category axis — work for a particular person, a particular type of difficulty.
And flags: backfilled, estimated, disputed. Its own note covers why those matter.
Projects specifically
A project field is worth having if you bill by project or estimate from past projects.
It should be a plain name, recorded consistently, with a start and end date recorded somewhere.
And it should not be the primary division, because in five years most of them mean nothing to you and the ones that do are better found by tag.
Querying across them
The pattern that works: filter by tag or project, total by category, over a date range.
If your structure supports that, it will answer most questions you have in a decade.
If it does not — if totals require picking through tags — the structure is wrong and it is worth fixing before the record gets longer.
What to check
Do you have both a category and tags, or only one structure doing both jobs?
How many distinct categories have you actually used?
Could you total your time by kind of activity across three years without double counting?
And do you tag anomalies — the ill days, the travel, the outages?
The point
Category, tag and project are three different structures.
Confusing them is the commonest structural mistake, and it is what makes old records hard to query.
Worth adding
Tags proliferate and that is tolerable, because tags are for finding rather than counting.
An annual tidy in the analysis rather than in the data keeps them usable.
Additionally
The pattern that works: filter by tag or project, total by category, over a date range.
If totals require picking through tags, the structure is wrong.
Finally
Tags are good for anomalies — ill, travelling, outage — and for flags like backfilled or estimated.
Those explain the shape of a period years later.
In summary
One category per entry from a short fixed list, any number of tags never used for totals, and project as its own field if you need it.
Three structures, three jobs.
More in this section
These notes are about keeping a personal record for years rather than for a fortnight: durability beats accuracy, and a rough record kept for five years answers questions a perfect record of one month cannot. Tool guides are included as practical comparisons, while the underlying method remains independent of any single product. Nothing here is legal, tax or medical advice: requirements differ by jurisdiction and situation.