Controlled vocabulary template for listing features
Build a controlled vocabulary template for listing features: preferred terms, variant terms, allowed values and spreadsheet columns that keep listings comparable.
Contents
- What a controlled vocabulary template does
- Choosing the spreadsheet columns
- Writing the preferred term list
- Recording variant terms as pointers
- Defining allowed values for each attribute
- Building the template step by step
- Normalizing feature vocabularies across sources
- Keeping the vocabulary under control
- Questions about controlled vocabulary templates
- What is a controlled vocabulary template?
- How is a vocabulary template different from a plain feature list?
- Where do variant terms go in the template?
- What are allowed values and why do they matter?
- Who should own and update the term list?
A controlled vocabulary template is a fixed set of columns and rules that turns free-text feature descriptions into one agreed term per concept. It records a preferred term for each feature, the variant terms that mean the same thing, and the allowed values a listing may use. The template lets you place details from different listings side by side without guessing what each label meant.
- A controlled vocabulary template fixes the column names and the allowed values before any listing is entered.
- Each concept gets one preferred term, with variant terms recorded as pointers to it rather than as separate entries.
- A feature list built on allowed values can be compared across sources without manual cleanup.
- The spreadsheet columns should separate the concept, its synonyms, its value type and any unit of measure.
- Someone must own the list, because unmanaged additions slowly turn a controlled vocabulary back into free text.
What a controlled vocabulary template does
A controlled vocabulary template is a reusable structure for describing features with a closed set of terms and values. It replaces free text with a small dictionary that every listing source must follow when it submits details. The result is a feature list in which the same concept carries the same label everywhere.
The template does two jobs at once. It defines the vocabulary itself, meaning the concepts and their preferred terms, and it defines the shape of the record, meaning the spreadsheet columns each entry must fill. Both parts matter, because a good term list stored in a loose table is still hard to merge.
A vocabulary template differs from a blank feature list. A blank list accepts whatever words a contributor types, while a controlled template accepts only entries that match an existing preferred term or a recorded variant term. That restriction is what makes side-by-side comparison possible.
Choosing the spreadsheet columns
The spreadsheet columns of a controlled vocabulary template should capture identity, meaning and permitted values for each entry. A practical layout gives every concept a stable identifier, a preferred term, a definition and a value type. Further columns record variant terms and the source that uses them.
Listing attributes such as size, material, color or power supply need a value type column. The value type states whether an attribute takes free text, a number, a date or a choice from a fixed set. Attributes that take a choice also need a separate table of allowed values.
Keep one row per concept and avoid merging several attributes into one column. A column that mixes material and finish forces every later reader to split the text again. The table below shows a workable skeleton for the main sheet.
| Column | Holds | Example entry |
|---|---|---|
| Concept ID | Stable code for the concept | C014 |
| Preferred term | The one approved label | Battery life |
| Definition | Plain sentence fixing the meaning | Time a full charge supports operation |
| Value type | Kind of value allowed | Number with unit |
| Variant terms | Other words mapped to the term | Runtime; operating time |
| Source | Listing feed that uses the term | Supplier sheet A |
Writing the preferred term list
A preferred term list gives each concept exactly one label that everyone agrees to use. The preferred term should be short, concrete and free of brand names or marketing wording. Write a definition beside it so that two readers reach the same meaning.
Choose terms that describe what the feature is rather than how it is sold. A label such as quiet operation carries judgment, while noise level carries a measurable attribute. Neutral wording keeps the vocabulary usable when listings change their marketing language.
Avoid near-duplicates in the preferred term list. If two terms describe the same attribute, one becomes preferred and the other moves to the variant terms. Review the list whenever a new source arrives, because a new feed often brings words the list does not yet cover.
Recording variant terms as pointers
Variant terms are the alternative words that a listing source uses for an existing concept. The template records each variant term as a pointer to the preferred term, never as a new entry in its own right. This keeps the list short while still capturing how suppliers phrase the same idea.
Store variant terms in a dedicated column or a separate mapping sheet with one variant per row. A single cell holding several synonyms separated by punctuation is hard to search and easy to break. One variant per row also lets you record which source used it.
Some variant terms are ambiguous and map to more than one concept. Log those separately and resolve them by reading the source listing, rather than assigning a default. A term that resists a single mapping may need a clearer definition on the preferred term.
Defining allowed values for each attribute
Allowed values are the closed set of entries that an attribute may take, and they turn a vocabulary into something a machine can check. An attribute such as power source might permit mains, battery or manual as its allowed values. Any submitted value outside that set is rejected or flagged for review.
Build the allowed values from the concepts in the preferred term list, not from the words in a single feed. Each allowed value needs its own short definition when the meaning is not obvious. Values that carry a unit, such as length or weight, should state the unit once for the attribute.
Keep allowed values stable and add new ones through a recorded decision. Silent edits to an allowed value break earlier comparisons without leaving a trace. The list below shows how one attribute and its permitted values fit together.
- Attribute: power source
- Allowed value: mains
- Allowed value: battery
- Allowed value: manual
- Value outside the set: flagged for owner review
Building the template step by step
Start the controlled vocabulary template from the listings you already hold, because real entries reveal the concepts you actually need. Collect the feature labels from each source and group the ones that share a meaning. Each group becomes a candidate concept for the preferred term list.
Next, assign a preferred term and a definition to each group and move the remaining labels into the variant terms. Set a value type for every concept and write the allowed values where the type is a fixed choice. Finally, test the template by entering a few listings and checking what fails.
Record the owner, the version and the change history alongside the template. A vocabulary without an owner drifts as contributors add local terms. Small, dated changes are easier to audit than one large rewrite later.
Normalizing feature vocabularies across sources
Normalizing feature vocabularies means reducing the different words from several listing sources to one shared set of preferred terms and allowed values. The template carries that mapping, so a reader can compare attributes without translating each source by hand. Differences in wording become visible and can be resolved once.
Run each incoming source against the variant terms before comparing any attributes. Values that map cleanly can be aligned in the shared columns, while unmapped values go to a review queue. This order prevents a stray synonym from appearing as a separate feature.
Keep the original source text in a separate column so that every mapped value can be traced back. Traceability matters when a source disputes a mapping or changes its wording. The normalized columns then support the shortlist, and the raw columns settle questions about what was actually submitted.
Keeping the vocabulary under control
A controlled vocabulary stays controlled only through a small set of rules that contributors can follow. Every new term needs a definition, a value type and an owner decision before it enters the list. Every variant term needs a pointer to an existing preferred term.
Review the template at fixed points in the workflow, such as when a new listing source joins. Check for duplicate concepts, undefined terms and values that bypass the allowed set. Record each change with a date and a reason so the history stays readable.
Treat the allowed values as part of the vocabulary and not as a footnote. Widening a value set quietly weakens comparison, while narrowing it too far rejects legitimate listings. Both moves deserve a written decision from the person who owns the list.
Questions about controlled vocabulary templates
What is a controlled vocabulary template?
A controlled vocabulary template is a fixed structure that holds one preferred term per feature, the variant terms that map to it, and the allowed values an entry may take. It combines a term list with defined spreadsheet columns, so every listing source submits details in the same shape. The template makes feature details comparable across sources.
How is a vocabulary template different from a plain feature list?
A vocabulary template restricts entries to a closed set of preferred terms and allowed values, while a plain feature list accepts any wording a contributor types. The restriction is what allows two listings to be aligned without manual rewriting. A plain feature list records what someone wrote, and the template records what the words are allowed to mean.
Where do variant terms go in the template?
Variant terms belong in a dedicated column or a separate mapping sheet, with one variant per row pointing to a preferred term. They are never added as new concepts, because that would duplicate the entry. Recording the source beside each variant helps when a feed changes its wording later.
What are allowed values and why do they matter?
Allowed values are the closed set of entries an attribute may take, such as mains, battery or manual for a power source. They let a reviewer or a script check each submitted value against the list. Values outside the set are flagged rather than silently accepted.
Who should own and update the term list?
One named owner should hold the term list and approve additions, because unmanaged edits turn a controlled vocabulary back into free text. That owner reviews new sources, checks for duplicate concepts and records each change with a date and a reason. Contributors can propose terms, but the owner decides what enters the list.