Seller OS
Fiverr tag research: how to choose better tags
Tag research means collecting candidate phrases before you open the tag field, then choosing the five that add search context your title does not already cover. Fiverr's help center says tags improve your gig's visibility in search and that you should use all five, so the real question is which five.
This guide covers where candidates come from, how tags differ from the positive keywords field, a worked example, and the one-tag-at-a-time test. Fiverr does not publish tag performance data, so every test here is judged by your own gig's impressions and orders.
Free plan available. Local-first data. Human review on every change.
Tags vs positive keywords: two different fields
Fiverr's help center describes tags as a visibility tool: you can add up to five tags to improve your gig's visibility in search, and the same guidance says to use all five, in Fiverr's Creating a Gig article. The positive keywords field is separate: up to five words or short phrases that help Fiverr categorize your service in the right context.
Read that as two pools of five phrases each, with different jobs. Positive keywords tell Fiverr what the service is; tags give the search system extra phrase variants buyers type. Keeping overlap low between the two pools, and between both pools and your title, gives you the widest coverage from the fields you control.
- Tag slot: a short phrase buyers might type, used for search visibility.
- Positive keyword: a phrase that describes what the service is.
- Title phrase: the anchor both fields should support, not repeat.
- The rules overview for the fields lives in Fiverr gig tags; this page is about researching them.
Where tag candidates actually come from
Build a candidate list of fifteen to twenty phrases before you touch the tag field. Your sources are the places buyer language already appears: Fiverr's search suggestions, the category and subcategory names around your service, the recurring phrases in the top gigs' titles and package names, and the wording buyers used in your own messages.
Your inbox is the most underused source. If three buyers asked about Shopify product page speed and your title says website speed optimization, that exact phrase is evidence for a tag. Fiverr's keyword data, where it exists, shows the search terms clients used, not which tag matched them, so inbox language is the closest free signal you get.
- Search suggestions: type a candidate into Fiverr search and note the completions.
- Category language: read your subcategory name and the ones next to it.
- Competitor copy: note phrases that repeat across titles and package names.
- Your inbox: collect the exact phrases buyers used when describing the job.
- The Fiverr keyword tool page covers the broader research workflow if you want it.
Worked example: five tags for a speed gig
Suppose your gig title is I will speed up your WordPress website. The primary phrase is already in the title, so tags should extend it rather than repeat it. The table below sorts a candidate list and keeps the strongest four, an illustration of the method rather than platform data.
Notice what the rejected candidates have in common: one repeats the title, and one belongs to a platform you do not serve. Repeating the title spends a slot without adding context; serving the wrong platform spends a slot on a search you cannot convert. The fifth slot goes to whichever remaining candidate has the strongest evidence, often a mobile variant.
| Candidate | Evidence | Decision |
|---|---|---|
| wordpress speed optimization | Already in the title | Reject: repeats the title |
| website speed optimization | Search suggestions | Keep: broader variant |
| core web vitals | Buyer messages | Keep: technical phrasing |
| site loading time | Competitor titles | Keep: plain-language variant |
| page speed audit | Package wording | Keep: buyer-stage phrasing |
| shopify speed optimization | Wrong platform for this gig | Reject: not your service |
Test tags one at a time, with a window
If you change three tags and impressions move, you have learned nothing about which tag did the work. Change one tag, leave the rest of the gig untouched, and let it run. Fiverr's own keyword guidance is a useful default window: give new or updated keywords at least 30 days before reviewing their impact.
A test needs a before and an after, so record the date, the tag you swapped, and your gig's impressions and orders for the prior period. The window is not a promise that a tag takes 30 days to work; it is a guardrail against judging a change on three days of noise.
Pick the tag
choose the candidate with the weakest evidence.
Record the baseline
note the current tag, the date, and last period's impressions and orders.
Make one swap
change the tag in the gig editor and leave every other field alone.
Wait the window
hold the full 30 days before reading the numbers.
Decide
keep the change only if the movement survives a second window.
“Give any new or updated keywords at least 30 days before reviewing their impact and analyzing their performance.”
— Fiverr Help Center, Advanced Analytics for Seller Plus, accessed September 2026
Tag mistakes that waste a slot
The most common mistake is treating tags as a place for adjectives: professional, fast, affordable. Buyers search for deliverables and problems, not praise. The second is duplicating your title word for word, which adds no new phrase to match. The third is spending all five slots on slight variants of one phrase, which narrows coverage instead of widening it.
One more: leaving slots empty because you are unsure. Fiverr's best-practice note says to use all five tags, and an empty slot cannot match anything. A weak candidate is still a test; one window tells you more than a guess does.
When to revisit tags and when to leave them
Revisit tags when something real changed: you added a service, dropped one, a buyer cohort shifted, or a test window closed with a clear result. Do not revisit because a week was quiet. One quiet week is noise, and each edit resets the test you were running.
Keep a short local log of tags and dates, two lines per test. It prevents the classic problem where you cannot remember which five tags were live when impressions moved, and it makes the next gig SEO audit faster to run.
Where Seller OS helps
Seller OS supports the research side of tags without pretending to know Fiverr's internal data. Keyword suggestions help you build the candidate list, and the gig audit reviews the fields a live gig already uses, including how title, tags, and description overlap. Everything runs locally in Chrome.
The honest limit: it cannot tell you which tag triggered a search, because Fiverr does not publish that. It keeps your candidate list and test log in one place, and on Pro the keyword monitor records movement while Chrome is open so a tag test has a timeline. You still make every edit and Save in Fiverr yourself.
Fiverr Tag Research questions
How many tags does Fiverr allow?
Fiverr's help center says you can add up to five tags to improve your gig's visibility in search, and its best-practice guidance is to use all five. Fiverr does not publish a tag character limit, so keep each tag a short, exact phrase rather than a sentence.
Are Fiverr tags the same as the positive keywords field?
No, they are two separate fields of up to five entries each. Fiverr describes positive keywords as phrases that help categorize your service, while tags are described as improving visibility in search. Treat them as two pools: one phrase's job is categorization, the other's is search coverage.
Can I see which tags other sellers use?
Fiverr does not publish tag data, so do not plan around copying another gig's tag list. Build candidates from what is public: search suggestions, category and subcategory names, recurring phrases in competitor titles and package names, and the exact wording buyers used in your own inbox and orders.
How long should I wait before judging a tag change?
Use the same guardrail Fiverr gives for keyword changes: at least 30 days before reviewing impact. Change one tag, leave the rest alone, and compare the next period against the one before it. A quiet week is not a verdict, and stacking several changes makes the result unreadable.
Give all five slots a job.
Build the candidate list, test one tag per window, and keep only the changes that move your gig.