Title optimisation is often reduced to “put more keywords in the title”. That is too shallow. A Shopping title has to do at least three jobs: accurately identify the product, expose the most useful differentiating information early and remain maintainable across hundreds or thousands of products.
Google's current title specification allows up to 150 characters and recommends putting the most important details first because users may see only the first 70 or fewer characters depending on the placement and device. Google also recommends using relevant product details and avoiding promotional language. That supports a simple principle: front-load identity and differentiation, not marketing noise.
Start with the product data specification, not an agency template
The title must accurately describe the product. It should not contain promotional text, unrelated search terms or information that belongs in other attributes. Google's optimisation guidance also encourages merchants to provide complete, accurate data and high-quality images alongside titles; title work is one part of catalogue quality, not a substitute for it.
That matters because a title built from unreliable data scales the unreliability. If colour is inconsistent, brand is wrong or variant capacity is missing, an automated title rule will repeat those errors across the catalogue with impressive efficiency.
Title optimisation starts one layer earlier
Before changing title order, check whether the attributes you want to use are complete, normalised and accurate. A title rule should assemble trusted components, not compensate for missing product structure.
Build titles from attributes, not manual copy
For a large catalogue, I prefer a title architecture built from fields such as brand, product name, product type, model, gender, colour, size or capacity. Which attributes belong in the title depends on what helps a shopper distinguish the item in that category.
A rule-based approach gives you three advantages:
- consistent structure across similar products;
- faster updates when source attributes change;
- the ability to test a title pattern by category rather than editing individual products.
Merchant Center attribute rules can combine incoming fields and static values, which makes them one possible implementation layer for channel-specific title construction. Feed-management platforms can do the same. The important thing is that the component data remains trustworthy and the logic is documented.
Different categories need different title logic
| Category | Potential priority information | Why |
|---|---|---|
| Branded fashion | Brand + product type/name + gender/audience + colour + size where useful | Brand and visible style are often key recognition signals. |
| Consumer electronics | Brand + model + product type + capacity/spec + colour | Model and technical specification can distinguish near-identical offers. |
| Furniture | Product name/type + key material + size/dimensions + colour | Physical fit and finish matter strongly to the decision. |
| Parts / B2B | Brand/manufacturer + part number + product type + compatibility detail where appropriate | Users often search with exact technical identifiers. |
| Private-label consumables | Brand + product type + core feature + quantity/size | Pack size and product type can be more useful than decorative naming. |
These are starting points, not universal formulas. Search behaviour, product imagery, category norms and the attributes actually available in your data should influence the final pattern.
Front-load the information that earns recognition
Because titles can be truncated, the first portion has disproportionate value. Avoid spending those characters on repetitive wording that is already obvious from the image or category if a more distinguishing attribute would help.
Weak
Amazing New Genuine Premium Trainers - Example Runner X Men's Shoes Black Size 9
Cleaner
Example Runner X Men's Trainers - Black - Size 9
The cleaner version identifies the product and variant quickly without promotional filler. Whether size belongs in the title depends on how offers are structured and how the title is displayed; the point is to use characters for information rather than adjectives.
Avoid duplicate information that adds no recognition
If every title begins “Buy online with free delivery from ExampleShop”, you have spent the most valuable part of the title on retailer promotion instead of product identity. Shipping and promotional information belong in the attributes and settings designed for them.
How I would test title changes
Do not rewrite the whole catalogue because one title formula sounds sensible. Treat title optimisation as a controlled data change.
Choose a coherent product group
Use one category or brand with enough traffic and consistent source data. Avoid mixing products whose search behaviour is completely different.
Define the hypothesis
For example: “Moving model and capacity earlier should improve relevance and click quality for this electronics range.”
Validate data coverage
If 35% of products have missing capacity, fix the attribute before making it a core part of the title rule.
Preview representative titles
Check short and long product names, special characters, variants and edge cases before deployment.
Measure the commercial outcome
Where traffic volume allows, compare visibility, clicks, CTR, conversion quality and downstream revenue rather than declaring success because titles look better.
Keep the rollback
Document the previous rule and change date so you can reverse or isolate the effect if performance moves unexpectedly.
A 2026 detail: Google distinguishes AI-generated titles
Google's current specification supports both title and structured_title. It states that titles created using generative AI should use the structured title attribute, which can indicate the digital source type. This is a good example of why feed architecture should follow the current specification rather than relying on old checklists.
That does not mean every retailer should generate titles with AI. Generative tooling can help at scale, but product accuracy, governance and human QA remain important. If an AI system invents a material, size or compatibility claim that is not present in source data, you have created a product-data problem rather than an optimisation.
Common title mistakes
- One template for every category: a furniture title and an electronics title often need different priorities.
- Keyword lists: titles should describe the product, not become a search-query dump.
- Promotional language: “best”, “free delivery”, sale messaging and retailer slogans waste title space and can conflict with requirements.
- Buried model/specification: the detail a shopper uses to distinguish the item appears too late.
- Rule built on dirty fields: inconsistent brand, colour or product type becomes visible at scale.
- No variant awareness: titles describe a parent product while the offer is for a specific size, colour or capacity.
- No measurement: the team celebrates a rewrite without knowing whether it improved relevant traffic or sales.
The title is an output of catalogue quality
The most scalable title programme I can build is one where the catalogue already contains accurate brand, product type, model and variant attributes. Then optimisation becomes controlled assembly rather than manual copywriting.
Common questions
How long can a Google Shopping product title be?
Google's current title specification allows up to 150 characters, but it notes that users often see only the first 70 or fewer depending on context. Put important details first.
Should I put keywords in product titles?
Use relevant words that accurately describe the product and match how customers recognise it. Do not turn the title into an unrelated keyword list.
Should brand always come first?
No universal rule fits every category. Brand is often important for recognised branded products, but the best ordering depends on category, intent and what distinguishes the item.
Can I use attribute rules to build titles?
Yes. Merchant Center attribute rules can combine incoming values and static text. The source fields should be accurate and the rule should be tested before wide deployment.
Official Google references reviewed for this guide
- Title and structured_title best practices
- Tips to optimise product data
- Attribute rules (formerly feed rules)
- Google Merchant Center product data specification
Merchant Center changes over time. This guide was reviewed against current Google documentation on 29 September 2026.
Want a second pair of eyes on the actual setup?
Start with the free Merchant Center & Feed Triage. No account access is needed for the first look. Send the website, catalogue size and the symptom and we will tell you where we would investigate first.