Instagram Hashtag Performance for Your Own Posts

Open the Hashtag Set Performance Analyzer with your own Instagram observations. The example below examines one pottery Reel series at a shared collection age.

Choose a Reel series with a recognizable content promise

Begin with a series you can describe without referring to its results. A pottery studio might review short demonstrations of centering clay, opening a form, and managing the rim, all aimed at beginners learning wheel basics. That definition gives the team a useful inclusion boundary before it sees which posts received more attention. Mixing those demonstrations with a shop announcement, a finished-product carousel, and an exhibition invitation would make the hashtag comparison harder to interpret. The analyzer can summarize those rows, but a higher median for one set could simply accompany a different creative purpose. Choose the publishing question first, then gather the records that actually belong to it.

For this guide, the fictional studio records eleven Reels from one account under the content topic Wheel basics and the format Reel. All measurements describe the posts at seventy-two hours after publication, and the declared traffic is organic. Those are assumptions of the teaching example, not claims that seventy-two hours is a universally sufficient collection window. A different studio can choose another age appropriate to its review as long as it collects the observations consistently. The tool does not infer post age from an ID or reduce lifetime totals to an earlier snapshot. Keep the publication and collection timestamps in your separate evidence sheet.

The practical question is whether two existing tag combinations are worth carrying into another comparable Reel batch, alongside a no-tag reference. It is not whether a pottery hashtag has a particular global search volume. Platform-wide interest, discovery placement, and audience behavior are different questions from the exact-set summary available here. Choose a primary outcome that fits the studio’s aim, and state what a useful follow-up would look like. For example, the team may prioritize interaction around teaching posts rather than raw exposure. That choice makes it less tempting to switch objectives afterward merely because a different column gives one set a more impressive position.

Record post measurements without borrowing account totals

Use the measurement source available for the specific post you manage, and confirm the meaning of each count before transcribing it. A total shown for the account over a date range is not automatically the result for one Reel. Likewise, a value that combines distribution contexts should not be treated as an Instagram-only observation without checking its scope. The analyzer has no connection that can detect these mistakes. Keep a source reference outside the project, along with the collection time and any relevant inclusion note. If the native screen changes its labels, resolve the definition rather than forcing the new figure into whichever old spreadsheet column looks closest.

The pottery example uses Reach and a consistently defined Engagements count to review actions per one hundred reach. The fictional studio defines that action total as the recorded likes, comments, saves, and shares for each Reel, under one consistent collection convention. This is an example definition, not an instruction to add every metric your interface happens to show. Do not mix an action total with unique accounts engaged, profile-level activity, or follower gains. If a component needed for your chosen definition is unavailable, either keep the entire total unknown or adopt a different consistently documented objective for a separate review.

For every saved row, enter its stable post ID, topic, format, exact tag list, and available counts. Leave unused measurement fields blank instead of copying the views number into reach or impressions. A known count of zero remains zero. An unavailable reach value remains blank even when engagements are known, because the ratio requires both. In this example one Reel has sixty-five recorded actions but unknown reach; it remains a tracked post without an eligible action-per-reach value. Keeping that distinction visible is more informative than inventing a denominator or dropping the record from the project so the set appears fully measured.

Capture the complete published set, not a convenient subset

The first four pottery Reels use #PotteryTips and #WheelThrowing. Four others use #PotteryTips and #ClayStudio. Three deliberately use no hashtags. These are the entire recorded sets for this example, not selected keywords extracted from a longer caption. If a real Reel also carried a campaign or location tag, include it in the row. Omitting that extra member would falsely merge a different published combination with a shorter set. Treat the list as evidence of what appeared on the post at the relevant observation, not as an editable recommendation field where the analyst can tidy away inconvenient details.

Case and order variations do not create additional groups. A row containing #wheelthrowing followed by #POTTERYTIPS belongs to the same normalized set as the first combination. Repeating #PotteryTips within a row contributes one distinct use, not two, and the tool reports repeated entries. Original tag text remains in the editable project even though grouping uses canonical keys. An underscore or an accent is meaningful, however: a differently spelled label is not silently mapped to a familiar English word. Review spelling differences against the source before deciding whether they are transcription errors or genuinely different labels used on the posts.

Blank tags are reserved for the deliberate no-tag observations. Do not use that state when you simply have not checked a caption yet; resolve the tag history before counting the post as a control. Paste only the tag list into the field, since it does not automatically extract hashtags from prose. Keep any caption edits in an external observation note because this project stores one recorded state per post, not a history of changes. If a Reel’s tags changed during collection, decide whether that row belongs in this comparison and document the decision without pretending the tool can reconstruct how long each version was visible.

Calculate the pottery example at the post level

For the four #PotteryTips #WheelThrowing Reels, enter reach and engagements as 800 and 40; 1,000 and 70; 1,200 and 60; and 1,400 and 112. Selecting Engagements per reach gives post values of 5, 7, 5, and 8 actions per one hundred reach. Their median is 6 and their mean is 6.25, with a range from 5 to 8. For #PotteryTips #ClayStudio, enter 900 and 54; 1,100 and 44; unknown reach and 65; and 1,300 and 78. Its eligible rates are 6, 4, and 6, with one unavailable rate. The median is also 6 and the mean is approximately 5.33.

The no-tag Reels have reach and engagements of 700 and 28; 1,000 and 50; and 1,200 and 72. Their post-level rates are 4, 5, and 6, so both their median and mean are 5. Across all eleven selected posts, ten rates are eligible and one is missing. The overall median is 5.5 actions per one hundred reach, and the overall mean is 5.6. There are three exact groups, including the blank set, and three distinct tag labels. At a minimum of three eligible posts, all three groups meet the review threshold, but the clay-studio group still has one missing measurement among its four recorded posts.

Both tagged groups receive a median index of approximately 109.09 against the selected baseline of 5.5; the no-tag group’s index is approximately 90.91. These are indices, not percentage-point changes in engagement rate or evidence of incremental reach. The tied medians also mean that table order must not be read as a winning recommendation: deterministic set-key ordering resolves the tie. The first group’s higher mean describes its supplied distribution, not proof that it is better. Do not pool all actions and reach to reproduce the displayed group mean, because the analyzer summarizes the individual eligible post ratios. Record the unit clearly when presenting these fictional results to a colleague.

Inspect missing reach before choosing a favorite

The sixty-five-action Reel with unknown reach illustrates a common interpretation problem. Its action total alone cannot establish whether its rate should be high or low. At a reach of 650 the ratio would be 10; at 1,300 it would be 5. Neither denominator is supplied in the example, so neither rate belongs in the analysis. Return to the authorized source to resolve the missing value when possible. Until then, describe the clay-studio group as three eligible rates from four posts. Do not call its unknown record a zero, and do not claim that meeting the minimum threshold makes the missing measurement irrelevant.

If the team’s actual question changes to reach per post, choose Reach as a separate objective and inspect eligibility again. In this fixture the first tagged group has a median reach of 1,100, the clay-studio group also has 1,100 from three known observations, and the no-tag group has 1,000. The overall eligible reach median is 1,050. That is a different summary from the action ratios and should be named accordingly. A post can support one objective while remaining ineligible for another. Save the chosen context with each review so a later comparison does not confuse an action-rate report with an exposure-count report.

Inspect content details before interpreting even a complete-looking table. A before-and-after reveal, a familiar face, a striking glaze, or a collaboration can change the appeal of a Reel within the same broad Wheel basics label. The format field does not verify equivalent creative. Use the observations to locate examples worth reviewing, then compare their teaching promise, duration, opening, offer, and distribution circumstances outside the analyzer. Correct genuine data-entry errors from evidence, but retain valid unusually high or low results. A transparent explanation of why a post differs is more valuable than a cleaner median created by removing every observation that challenges the preferred story.

Read reuse and overlap as a record of your publishing

In the eleven-Reel example, #PotteryTips appears on eight posts, or approximately 72.73 percent of the selected dataset. Seven of those posts have eligible rates and one does not. #WheelThrowing appears on four posts with four eligible rates; #ClayStudio appears on four with three eligible rates and one missing. The shared #PotteryTips label occurs in two exact sets, while each other label occurs in one. These counts describe the studio’s recorded usage, not Instagram search demand, hashtag followers, or the size of an audience reached through a tag. The unknown denominator on one Reel does not erase the fact that its tag was used.

Each tagged pair occurs on four selected posts, giving approximately 36.36 percent co-use share for each pair. The two nonempty sets share #PotteryTips and have three different labels in their combined membership, so their overlap is approximately 33.33 percent. The no-tag group does not enter the overlap table. This tells the studio exactly what changed between the two combinations: #WheelThrowing was exchanged for #ClayStudio while #PotteryTips remained. It does not tell the studio which label caused an interaction, and it does not measure similarity between the audiences of those Reels. Keep set-membership evidence separate from performance evidence even when both appear in the same report.

Use that membership view to avoid an accidental multi-part revision. Adding a local tag, replacing the niche tag, and appending a campaign label would create a new set with several changed members, making the practical comparison less focused. A clear next note might say that the two existing complete sets will be reviewed again on the next wheel-basics batch, without changing the definition of engagements. Copy a set only after checking that it suits the actual upcoming post; reuse is not an automatic relevance check. For a deeper explanation of the token calculation, read how exact hashtag sets overlap.

Plan the next batch without rewriting the first one

Decide the next collection procedure before publishing another Reel. Specify the account, teaching series, format, traffic scope, observation age, primary objective, exact sets, and how no-tag posts will be included. Spread the candidate sets across comparable production circumstances instead of reserving one for unusually strong demonstrations or a launch week. A planned rotation can make the record easier to interpret, although it does not randomize who encounters each organic post. Keep that limitation in the conclusion rather than treating a neat schedule as proof of equal audiences. The purpose is a more informative local review, not a claim that the surrounding distribution system has been controlled.

Keep the earlier project intact and start a separately named batch when the conditions materially change. A later batch may have different seasonal interest, follower composition, or production quality even when the tag sets are identical. Review it on its own before combining observations, and explain why a combined summary is appropriate if you eventually make one. Do not keep adding posts until a preferred set moves to the top and then describe that stopping point as planned. Use the hashtag testing plan to write a finite collection and review routine. The analyzer’s minimum-post setting does not choose that routine for you.

At review time, reconcile every expected post against the saved rows, then resolve missing measurements without guessing. Read medians, means, ranges, and eligibility before choosing the next operational action. A tie can justify keeping both combinations in consideration; an inconclusive batch does not require a forced winner. The studio may instead discover that its measurement collection needs improvement or that the set difference is less important than a clearer teaching opening. Write that finding plainly. The free set analyzer supports this evidence review, but the decision should retain the creative context and collection details that no compact statistics table can fully express.

Save a useful review and a separate publishing brief

Save the editable project JSON with a recognizable batch name before closing the browser. It preserves the supplied rows, entered tags, settings, filters, and next-test note so you can reopen the working review. For a batch import, adapt the analyzer’s exact nine-column CSV template rather than assuming an Instagram export can be opened unchanged. The report CSV is a review output, not that input template and not the editable backup. A small source ledger outside the tool should retain collection timestamps and measurement references, allowing another reviewer to reconcile the eleven saved rows with the actual observations without asking the analyzer to retrieve them.

Before sharing, remember that filtering the review does not remove other saved posts from the export. Copy results, report CSV, and PDF carry all raw source rows and context, including the next-test note; the PDF also attaches the full project JSON and report CSV. A concise visible summary can therefore include private material in its attachments. For a client-facing discussion, prepare a separate version containing only appropriate rows and neutral account labels, regenerate it, and inspect the full file. The browser-only calculation does not make a downloaded document anonymous, and this analyzer does not offer a summary-only privacy mode that hides those underlying records.

Turn the chosen follow-up into a short publishing brief: the teaching promise, exact set, observation age, and person responsible for collecting results. For supported Instagram account routines, explore SMTasker; Skye and permissioned MCP can assist with available workspace actions. That is separate from this manually supplied review, not a connection to Reel insights or an automatic experiment import. Keep the measurement owner accountable for the next snapshot. Return to all free tools for other preparation needs, and keep the useful conclusion modest: these recorded sets have been reviewed in a defined Reel series, with the remaining uncertainties visible.

Questions about Instagram Hashtag Performance

Continue with another tag-set guide

Review your supplied post observations in the Hashtag Set Performance Analyzer, or choose another guide below. Find more practical resources in all free tools.