# Frequently asked questions

Quick answers about Starlecta: the data, the AI analysis, the scores and how to get access.

## About Starlecta

### What is Starlecta?

Starlecta is the live feed of the influencer marketing ecosystem. It tracks influencers on Instagram, shows which brands they work with and how their content performs, and turns it into reports your team can act on.

### Who is Starlecta for?

Brands, agencies, e-commerce marketplaces and marketing teams that plan, run and measure influencer campaigns.

### What can I do with Starlecta?

Find the right influencers with AI Search, compare them side by side, follow your competitors' influencer activity, track Stories, Reels and posts with their links, and get AI-written reports on any brand or sector.

### Which markets do you cover?

Starlecta is built as a global product and starts with Turkey, where it tracks the active Instagram influencer market. New markets are on the way.

## Data and coverage

### How much data does Starlecta track?

More than 53,000 Instagram influencer accounts, 9,800+ brands and 1,500+ shopping platforms and sites, with around 200,000 new Stories, Reels and posts a day.

### Which content do you cover?

Instagram Stories, Reels and posts, together with the links, products, brands and platforms inside them.

### How often is the data updated?

Every day. Stories and content are collected daily, scores are refreshed daily, and the trending hashtags in Trend Cloud refresh every hour.

### How do you detect a brand collaboration?

From the content itself: brand mentions and tags, ad disclosures, shopping links and the products behind them. Each collaboration is linked to its content, so you can always see the proof.

### Can an influencer ask to be removed?

Yes. Write to support@starlecta.com and we remove the account from the public pages and tools.

### What does link analysis show?

Which shopping platforms, brands and products influencers send their followers to, and how that changes day by day.

## AI analysis and reports

### What does AI Intel do?

AI Intel watches influencer videos for you: it transcribes the speech, reads what is on screen and puts your competitors' influencer strategy in front of you as a clear report.

### What does the AI analyse in a report?

Every Reel, story and post is analysed: videos are watched and transcribed, images and captions are read. The AI then reads the creative and its format, the language and tone, the opening hook, the campaign and its message, the ad strategy behind it, the slogans and mottos that repeat, and the insights a brand team can act on.

### How does AI Search work?

Write what you are looking for in one sentence, in Turkish or English. AI Search turns it into filters and returns a ranked shortlist of matching influencers.

### What is in a sector report?

The leading brands of a sector for the week, how much influencer content each one has, which formats and influencers carry it, the standout content and the AI analysis of the campaigns.

### Can I get a report for my own brand?

Yes. Book a demo with your brand and competitors and we prepare it for you.

## Starlecta Quality Score (SQS)

### What is the Starlecta Quality Score?

A 0–100 score that shows how strong an influencer account is as a partner, in one number.

### What goes into the Starlecta Quality Score?

Engagement, audience quality and authenticity, content consistency and how the account performs against similar accounts.

## Engagement rate

### How is the engagement rate calculated?

Likes plus comments on recent posts, divided by followers, as a percentage. Reels also get their own rate based on views.

### What is a good engagement rate?

It depends on the account size and category, so Starlecta always compares an influencer with accounts of a similar size in the same category.

## Audience analysis

### What does the audience analysis show?

Who follows an influencer: gender, age, city and interests, and how real and engaged that audience is.

### Can I see audience overlap between influencers?

Yes. Audience Overlap shows how much of two audiences is shared, so you do not pay twice to reach the same people.

## Story views

### How do you estimate Story views?

From each influencer's own Reel views and audience, using a model built on similar accounts. You get a clear range you can plan with.

## Access and tools

### Which tools can I try for free?

Influencer profile, Influencer compare and Influencer AI Search run on our data with no sign-up, plus the engagement rate calculator, Story views estimator, campaign planner and performance benchmark.

### How do I get full access?

Book a demo. We show you the panel with your own brand and competitors in a 30-minute call and set up your account.

## About Starlecta

### What is Starlecta?

Starlecta is an influencer analytics platform for brands and agencies: track rival brands' Instagram influencer collaborations — Stories, posts and Reels — compare influencers and prepare reports with stated sources and methods.

### Who is Starlecta for?

Brand teams and agencies that want to see which influencers rival brands work with, compare influencers before a campaign and prepare reports for clients. The sample reports are public.

### What can I do with Starlecta?

Find influencers with AI Search from one sentence, compare two influencers on the same dates and method, look up the open card of an influencer, follow brand collaborations in Stories, posts and Reels, and read the weekly sector reports. The three tools run on tracked data with no sign-up; the full panel opens in a demo.

### Which markets do you cover?

Starlecta is built as a global product. Today it tracks influencers in one market, Turkey, and every figure on this site is counted on those tracked accounts. Global and further markets are coming; until they are live no figure here claims to cover them.

## Data coverage and counting

### Which platforms and content does Starlecta cover?

Instagram: Stories, posts and Reels from the accounts Starlecta tracks. WhatsApp and Telegram are out of scope for brand-collaboration counts.

### How often is the data updated?

Every figure carries its date. The quality score is calculated once a day and stored. The live pulse refreshes hourly for the last day and nightly for the last 7 days. Weekly sector reports state the period and the date of their data.

### How is a brand collaboration verified?

Each report states its own rule: a post counts when it carries a verified brand link, or caption or tag evidence of the brand. A tag or mention alone is not proof of a paid collaboration.

### How are contents counted?

Starlecta counts content, not links or mentions. One registered content record is one unit, however many links, platforms or brand matches it carries.

### What is the counting unit?

- The unit is the registered content record: one post, Reel or Story. In the brand view the count is the number of distinct content records behind the matched links, so a content with five links counts once.
- Accounts are counted as distinct accounts, separately from contents.
- Weekly volume counts distinct contents per Monday-start week in Turkey time.

### How is content with several platforms or brands counted?

- A content linked to two platforms, or naming two compared brands, appears in both rows. Row counts must therefore not be added up.
- The unique total under the rows is the distinct count across the whole column. It is never the sum of the rows.
- A share table states its basis. Network-linked shares use the sum of each network's unique contents as the denominator, while the influencer total is deduplicated across networks.
- In the sample reports a post naming more than one brand is counted once per brand, so the per-brand counts can add up to more than the distinct items.

### How are co-authored posts and re-captured Stories counted?

- The sample-report pipeline deduplicates a co-authored post shared by two accounts into one content. The panel counts distinct content records, and the collector deduplicates on the account, the provider content id and the posting time, so the same provider post on two accounts can be two records in the panel.
- A Story captured again can land as a second record. The collector checks the account and the provider content id before saving to prevent it, and the counting layer does not repeat that check.

### How are near-duplicate creatives counted?

- For visual analysis, images that repeat across accounts are grouped as one creative when their perceptual hashes are within a distance of 6, separately for each brand and content kind.
- The report states how many unique creatives were analysed. This grouping serves the creative analysis only and does not change the content counts above.

### What are the scope, period and limits of the counts?

- Counts cover Starlecta's tracked accounts and the selected period. Absolute volumes are lower bounds, because the number of tracked accounts grew over time.
- Counting once is not a test of originality: the same media posted by two accounts is two contents.
- The brand view counts contents with a registered link match; caption, mention and hashtag evidence are not tracked there. The sample reports also use caption and tag evidence and say so.
- Observed sharing is not a confirmed or paid collaboration.

## Figures and sources

### How many accounts does the platform track?

53,000+ accounts tracked. Instagram influencer accounts we track. Source: Starlecta database, 29 October 2026.

### How many tracked accounts post in a week?

41,000+ accounts active weekly. Accounts active weekly = tracked accounts that posted in the last 7 completed days (21–27 October 2026): 41,249 (Patterns, 28 October 2026).

### How many shares, Stories and links are tracked?

10.8M+ shares tracked; 7.7M+ Stories tracked; 1.6M+ links tracked in shares; Almost 1 in 5 Stories carry a link. Shares, Stories and links in shares = Patterns totals, 28 October 2026: 10.81M shares, 7.73M Stories, 1.61M links. Almost 1 in 5 Stories carry a link = 124,219 of 657,393 Stories posted 21–27 October 2026 (18.9%).

### How many audience profiles are stored?

54M+ audience profiles. Audience profile records in our database; not reach. Source: Starlecta database, 29 October 2026. Audience profiles = audience profile records in the Starlecta database, 29 October 2026; not reach.

### How many contents are posted in a day?

About 200K contents a day. Stories, Reels and posts, by posting day. Contents a day = Instagram Stories, Reels and posts recorded per posting day (TRT); measured 26–27 October 2026, average 199,757.

### How many comments are stored?

486M+ comments. Comments stored in our database. Comments = the number of comment rows stored in the Starlecta database, 29 October 2026.

### How many brands does the catalogue hold?

9,800+. Brands in our catalogue. Counted 30 October 2026, rounded down.

### How many brands were active in the last 30 days?

4,600+. Brands seen in influencer links in the last 30 days. Counted 30 October 2026, rounded down.

### How many platforms and sites appear in influencer links?

1,500+. Shopping platforms and sites seen in links, last 30 days. Counted 30 October 2026, rounded down.

### How many products were linked in the last 30 days?

18,000+. Distinct products linked in the last 30 days. Counted 30 October 2026, rounded down.

### How many links were shared in the last 30 days?

400K+ links shared in 30 days. Counted 30 October 2026, rounded down.

## Starlecta Quality Score (SQS)

### What is the Starlecta Quality Score?

SQS is a 0 to 100 score for how well an influencer performs. It is calculated once a day and stored; reports and the panel read the stored value and never recompute it.

### What is the Starlecta Quality Score made of?

The score is the sum of seven components, each worth up to its weight in points.

- Engagement rate: 25
- Story activity: 20 (how many Stories in the last 30 days)
- Real audience: 15 (share of analysed commenters counted as real people)
- Recency: 15
- Loyal commenters: 10 (share who comment again)
- Growth: 10 (follower growth over 30 days)
- Positive sentiment: 5 (share of comments with positive tone)

### How is each component scored?

- Most components are an exact percentile rank against peers. For engagement, Story activity and sentiment the peers are the same category first; when the category is too thin, the account's own follower tier. A thin category is never merged with another one.
- Real audience, loyal commenters and growth are ranked against the follower tier.
- Recency is absolute, not a rank: the newer of the last post or Reel (seven days of grace, then a gradual decay) and the last Story (one day of grace, then a faster decay). An account with nothing on record scores at a small floor.
- The engagement rank is additionally scaled by how many recent feed items the rate rests on: with few items, an above-average rank shrinks, and with too few the rate is treated as unmeasured.

### What happens when a component cannot be measured?

- A component that was never measured counts as a neutral half and its weight is removed from coverage. The weights are not redistributed onto the components that did resolve.
- An observed zero is not missing: an account that posted no Stories has been measured at zero and scores that way.
- Coverage is the share of total weight that was measured. Confidence is high at 85% and above, medium at 65% and above, low at 40% and above.
- Below 40% coverage no score is shown.

### What does the Starlecta Quality Score not show?

- It is not a conversion or sales prediction, and not a fake-follower percentage.
- Commercial evidence (how often an account was paid for) is not part of the score; the score is about how well an influencer performs.
- The audience components rest on the accounts that comment, not on the whole follower base.

## Engagement rate

### What is the engagement rate?

Engagement rate is the share of followers who like or comment on an influencer's content. Starlecta stores two rates: an overall rate over posts and Reels, and a Reel-only rate.

### How is the engagement rate calculated?

- Per item: (likes + comments) divided by the account's current follower count, as a percent.
- The window is the last 90 days of posts and Reels. Newer items weigh more: 1.0 for 0 to 30 days, 0.6 for 30 to 60 days and 0.3 for 60 to 90 days.
- The account rate is the weighted median of the item rates, so a single viral post does not move it much.
- Play counts take no part in the rate. Reels and posts are on the same follower denominator.

### What is the difference between the overall and the Reels rate?

- Overall: posts and Reels together. This is the rate that is scored in SQS, sorted, filtered and benchmarked.
- Reel: Reels only, shown on profiles and never scored.
- The two are never used interchangeably.

### Why do some profiles say not measurable?

- An item whose like count is hidden or missing is left out, because comments alone would read as near zero. An authoritative zero stays.
- A rate is published only with enough measurable items (an effective sample of at least 3). Below that the value shows as "--", not as zero.
- Stories are not part of the engagement rate.

### How do I read the engagement rate?

- The denominator is the current follower count, not the count on the day of the post.
- It measures reaction relative to audience size. It is not reach, and it does not say whether the reactions came from real people; see the real audience component of SQS.

## Audience analysis

### What does the audience analysis show?

Starlecta describes an audience from the accounts that comment on an influencer's content, not from the whole follower list. The scan is stored per account; reports and the panel read the stored values.

### What is analysed in an audience profile?

- The source is the engaged commenters of the account's content. The follower list is not scanned. In the influencer report, comments are capped at 1,000 per content item.
- Every percentage describes that commenter sample, whose size is stored with the profile. Audience values are labelled estimated and are never presented as a share of all followers.

### Which audience fields are stored?

- Splits by age, gender, city, country, language and account type (personal or business).
- Interests, inferred from the commenters' engagement network. They are signals, not a share of people.
- Real audience (an authenticity signal in the engaging commenter profiles), loyal commenters (the share who comment again) and the share of commenters outside the country.
- Sentiment: the positive, neutral and negative tone of the classified comments, with the number of classified comments.
- The size of the commenter sample, and the margin of error of the age split.

### When is an audience profile shown?

- A profile needs a completed scan with a real-audience ratio and at least 30 analysed commenters. Below that the surface shows a dash or an empty state, not a precise-looking percentage.
- A split is also withheld when it is empty or when one bucket holds essentially all the weight, which is what a tiny sample produces.
- A value that cannot be measured is shown as not measurable. It is never shown as zero.

### How is the audience of a group of influencers calculated?

- For a set of influencers, such as the influencers who shared a brand in the selected period, the profile is the unweighted mean of their individual profiles. Every influencer counts equally, whatever their reach.
- Only influencers who pass the profile gate are included. Each dimension (city, age, gender, interests, country) is averaged over the influencers that have data for it, and a missing dimension is never counted as 0%.
- The view shows the top values of each dimension and states its coverage as n of N: the influencers with an evaluable profile out of all in the set. The label says it is a commenter sample, not followers.
- A report may show a second, visibility-weighted figure beside the equal-weight one; it is labelled as such.
- Audience similarity compares the measurable audience profiles of two influencers. Shared engaged people are a commenter overlap. Neither is shared followers.

### What are the scope, period and limits of audience data?

- Scope: Starlecta's tracked accounts. A profile is the latest stored scan of the account; it is not recomputed for the report period, and the sample reports state their data date.
- A group profile covers the influencers who shared in the selected period, each with the profile stored for that account.
- Commenters are more active than the typical follower, so the profile is not a census of the follower base.
- It does not identify people and is not a fake-follower percentage. The real audience, loyal commenters and positive sentiment components of the SQS read these same stored values.

## Story views estimate

### How are Instagram Story views handled?

The externally observed data Starlecta can access holds no measured view counts for Stories. Story views are therefore shown as a modelled estimate, labelled as an estimate and kept separate from measured Reel and post numbers.

### What is the Story views estimate?

Stories last 24 hours and the externally observed data we can access holds no measured view counts for them. Starlecta therefore shows Story views as a modelled estimate, labelled as an estimate, and keeps it apart from measured Reel and post numbers.

### How are Story views estimated?

- Estimated views = followers x a view ratio x the number of verified Stories.
- The view ratio is the median of the account's own Reel plays divided by followers, when it has at least three Reels with usable counts. That median is pulled into the p10 to p90 band of comparable peers, and the ratio never exceeds 1.
- When the account's own Reel sample is short, the median of peers is used: same category and follower tier when that group has at least 100 accounts, otherwise the follower tier.
- A low and a high scenario use the p10 and p90 ratios of the peer group. This range is a modelled spread, not a confidence interval and not measured past views.

### How is Story interaction modelled?

- Modelled interactions = estimated Story views x the account's median of (likes + comments) / plays on its Reels. At least three eligible Reels are needed.
- Reels with hidden or missing counters are left out of that median; a measured zero is kept.
- The overall engagement rate does not enter this calculation.

### How do I read the Story estimate?

- It is an estimate. It is not a measured view count, not unique reach and not a count of people.
- Measured and estimated values are never added into one number. A report shows Story estimates and Reel or post counters in separate lines.
- When a follower count or the needed peer group is missing, the estimate is left empty with a reason. It is not shown as zero.
- In the influencer report the estimate can come from the peer median, not from the account's own Reels; the report says so.

## Access, tools and reports

### How can I get access?

Ask for a demo at sales@starlecta.com and we show the platform on tracked data. The sample reports and method pages are public and need no sign-up.

### Which tools are free and do they need a sign-up?

Influencer lookup, Compare two influencers and Influencer search (with AI Search) run on our tracked data. They need no sign-up; each shows only the open part of an influencer card and counts the rest.

### Which reports do you publish?

Weekly sector reports and the four sample reports are public. A report page shows its first metrics and a brief; the PDF is the full report. Influencer names in public reports are anonymised.

### Can I get a report for my own brand?

Yes. Book a demo with your brand and competitor set; we prepare the report on tracked data and state its sources and method.

- [Docs: methods, guides and questions](https://starlecta.com/en/docs.md)
- [Try it](https://starlecta.com/en/demo.md)
- [Book a demo](https://starlecta.com/en/demo-request)

- **Definitions as of:** 24 October 2026

© 2026 Starlecta. All rights reserved. Content, data and reports on this site may not be copied or redistributed without written permission.
