TikTok plus Instagram: Cross-Platform Intelligence Patterns

If you only watch TikTok or only watch Instagram, you are reading a half-finished book. Creators publish on both. Audiences scroll on both. Campaign budgets land on both. Yet most analytics stacks still bolt a TikTok scraper to one dashboard and a Meta Graph integration to another, never reconciling them. The result is double-counted reach, mismatched creator names, and attribution that quietly favors whichever platform reports first.

This post is for marketing ops engineers, growth marketers, and agency platform leads building the joined view themselves. We will be honest up front: TikLiveAPI only covers TikTok. It does not call Instagram, Meta, or any other network. The interesting work is gluing it to Meta's official APIs so the two sides feed one normalized creator dimension table. That is the pattern this post describes end-to-end.

Why a single-platform view is incomplete

Single-platform thinking distorts three decisions that marketing ops makes every week.

Reach is double-counted by default

A creator with 500K TikTok followers and 300K Instagram followers does not have 800K reachable people. Audience overlap on lifestyle and beauty creators routinely sits between 20% and 45%. Without a joined identity, briefs over-promise impressions and CPMs look artificially cheap.

Content cadence is misread

A creator who posts daily Reels but only weekly TikToks looks "inactive" in a TikTok-only dashboard. Their actual content velocity is high, just weighted toward Meta. Cross-platform last_post_time fixes this.

Attribution drifts to whoever measures loudest

Meta Pixel and TikTok Pixel both claim the same conversion. Without a shared creator dimension and a deterministic dedupe rule, your monthly report becomes a negotiation.

The cross-platform signals worth tracking

Four signals consistently pay for the engineering effort.

1. Creator presence on both platforms. Pull TikTok bio_link via TikLiveAPI's /userinfo-by-username/; pull Instagram bio "website" via Meta's Instagram Graph API for Business. When either points at the other handle, you have a high-confidence match.
2. Simultaneous content posting. Many creators cross-post Reels to TikTok within the same hour. Comparing create_time from TikLiveAPI's /user-posts/ against Instagram media timestamp tells you whether they treat the platforms as twins or as separate creative tracks.
3. Audience overlap estimation. You cannot get user-level overlap (neither platform exposes it), but you can estimate via shared follower handles in public follower lists, geo/language distribution comparison, and engagement-rate symmetry.
4. Ad attribution across both. Once you have a canonical creator id, you can attach Meta Ads spend and TikTok Ads spend to the same row and compute combined CAC honestly.

Architecture: two providers, one creator dimension

Here is the text architecture diagram. The whole point of the design is that nothing downstream of the normalization layer knows which platform a row came from.

+-------------------------+        +----------------------------------+
|  TikTok side            |        |  Instagram side                  |
|                         |        |                                  |
|  TikLiveAPI             |        |  Meta Graph API (Business)       |
|  X-Api-Key auth         |        |  OAuth2 access tokens            |
|                         |        |                                  |
|  GET /userinfo-by-      |        |  /{ig-user-id}?fields=...        |
|      username/          |        |  /{ig-user-id}/media             |
|  GET /userinfo-by-id/   |        |                                  |
|  GET /user-posts/       |        |  Instagram Basic Display API     |
|  GET /user-followers/   |        |  (creator-owned, lightweight)    |
|  GET /search-user/      |        |                                  |
+-----------+-------------+        +------------------+---------------+
            |                                         |
            v                                         v
   +--------+-------------+              +------------+--------+
   | tiktok_raw (JSON)    |              | ig_raw (JSON)       |
   | partitioned by day   |              | partitioned by day  |
   +--------+-------------+              +------------+--------+
            |                                         |
            +--------------------+--------------------+
                                 v
                  +--------------+--------------+
                  | match_creators job          |
                  | (bio_link + username +      |
                  |  manual review queue)       |
                  +--------------+--------------+
                                 v
                  +--------------+--------------+
                  | cross_platform_creators     |
                  | canonical creator dimension |
                  +--------------+--------------+
                                 v
                  +--------------+--------------+
                  | downstream: briefs, CRM,    |
                  | BI, ads attribution         |
                  +-----------------------------+

The shape matters: raw tables stay platform-specific (so you never lie about the provenance of a field), and a single match job is the only place that decides "this TikTok account and this Instagram account are the same human."

Matching creators across platforms

This is the hardest engineering problem in the pipeline. Three matching rules, applied in order, get you most of the way.

Rule 1: bio_link cross-reference (highest confidence)

From TikLiveAPI's /userinfo-by-username/ the user.bioLink field carries the creator's external URL. If it contains instagram.com/{handle}, you have a one-way attestation from the creator themselves. Do the symmetric check from Instagram's website field. Both directions matching is the gold standard.

Rule 2: exact handle match (medium confidence)

Roughly 60% of mid-tier creators use the same handle on both platforms. A normalized lowercase comparison of user.uniqueId against the Instagram username catches them. Flag, do not auto-merge - "marie" on TikTok and "marie" on Instagram are not necessarily the same person.

Rule 3: manual review queue

For everything else, route to a human. The queue should show: avatar, follower count, last 3 post thumbnails, bio text. A trained reviewer can resolve 80 to 120 candidates per hour. Store the decision with a reviewer id and timestamp so you can audit later.

Pull the TikTok side of a candidate match with one call to /documentation/users/user-info/ using the canonical username, or by id when you already have it. For discovery from a search term, /documentation/users/search/ is the right entry point.

The data model

Keep the canonical table small. Heavy fields live in raw tables; the dimension is just the spine.

CREATE TABLE cross_platform_creators (
  creator_id          BIGINT PRIMARY KEY AUTO_INCREMENT,
  display_name        VARCHAR(255) NOT NULL,
  tiktok_username     VARCHAR(255) NULL,
  tiktok_user_id      VARCHAR(64)  NULL,
  tiktok_sec_uid      VARCHAR(128) NULL,
  instagram_handle    VARCHAR(255) NULL,
  instagram_ig_id     VARCHAR(64)  NULL,
  tiktok_followers    INT          NULL,
  instagram_followers INT          NULL,
  tiktok_last_post    DATETIME     NULL,
  instagram_last_post DATETIME     NULL,
  match_method        ENUM('bio_link','username','manual') NOT NULL,
  match_confidence    DECIMAL(3,2) NOT NULL,
  reviewed_by         INT          NULL,
  reviewed_at         DATETIME     NULL,
  created_at          DATETIME     NOT NULL,
  updated_at          DATETIME     NOT NULL,
  UNIQUE KEY uq_tt (tiktok_user_id),
  UNIQUE KEY uq_ig (instagram_ig_id),
  KEY idx_handles (tiktok_username, instagram_handle)
);

Notes on the shape:

• tiktok_user_id and instagram_ig_id are the durable keys. Handles change; ids do not.
• Store tiktok_sec_uid because some TikLiveAPI endpoints prefer it.
• Follower counts live here as a snapshot for fast joins; the history of counts belongs in a separate fact table.
• match_confidence lets you filter "only show bio_link-verified creators" in briefs without changing schema.

Computing combined reach and overlap

With the dimension in place, combined reach is one query. Honest combined reach subtracts an overlap estimate.

SELECT
  c.creator_id,
  c.display_name,
  c.tiktok_followers,
  c.instagram_followers,
  c.tiktok_followers + c.instagram_followers AS naive_reach,
  ROUND(
    (c.tiktok_followers + c.instagram_followers)
    * (1 - o.overlap_estimate),
    0
  ) AS deduped_reach
FROM cross_platform_creators c
LEFT JOIN audience_overlap_estimates o
  ON o.creator_id = c.creator_id
WHERE c.match_confidence >= 0.80;

For overlap estimation, the most defensible cheap method is:

1. Pull a sample of the TikTok follower list via /user-followers/ (cursor through ~1,000 entries).
2. For each follower with a bio_link, check whether that link points to an Instagram handle you can find in the same creator's Instagram followers (via Meta Graph API for accounts you have access to) or in the cross_platform_creators table for known matches.
3. Compute observed overlap on the sample, then apply a conservative correction factor.

This will not give you Nielsen-grade numbers. It will give you a defensible directional estimate, which is what briefs actually need.

Case studies for the joined data

Multi-platform campaign briefs

An agency planning a product launch wants 5M deduped impressions across 12 creators. Without the joined view, they pick creators based on TikTok alone and over-deliver on Instagram by 30% (wasted budget) or under-deliver (missed KPI). With the joined view, the brief is built from deduped_reach directly.

Full-funnel creator marketing

Map TikTok to top-of-funnel (entertainment, discovery) and Instagram Stories/Reels to mid-funnel (consideration). The cross_platform_creators table lets a planner say "this creator does both well, give them the full-funnel package" versus "TikTok only, pair with a different IG creator."

Holistic competitor monitoring

Tracking a competitor brand's TikTok via TikLiveAPI's /user-posts/ tells you half the story. Pairing it with their Instagram /media endpoint gives you the full publishing calendar. Plot post times together and you can often see their content team's working hours.

Production patterns

Parallel polling per platform

Run two independent worker pools, one per platform. They share nothing except writes to raw tables. This isolates failure: a Meta rate-limit pause should never block TikTok ingestion. Use a job queue with per-platform concurrency limits.

Dedupe by canonical creator id

Every downstream consumer joins on creator_id, never on handle. This is the single rule that prevents the "marie" collision problem from leaking into reports.

Snapshot on a schedule, do not poll on read

Refresh follower counts and last_post_time once or twice a day per creator. Reports read from the snapshot. This keeps API costs predictable and reports fast.

Backfill with idempotent keys

When you discover a new creator, backfill their posts in pages using the cursor + hasMore contract from /user-posts/. Key each row by aweme_id so reruns are safe.

Compliance notes for each platform's ToS

This is the slide your legal team will ask for.

• TikTok. TikLiveAPI surfaces publicly available data. You are still responsible for how you store and use it. Do not republish full user profiles; do not build derivative products that look like a TikTok clone.
• Instagram / Meta. Use the official Meta Graph API for Business with proper app review for any production use beyond your own owned accounts. The legacy Instagram Basic Display API is creator-token-scoped and not suitable for agency-side aggregation across many creators you do not own.
• Both. Honor deletions. If a creator deletes a post on either platform, your snapshot should mark it deleted within your refresh window. Reporting on deleted content is the fastest way to a takedown notice.
• PII. Follower lists are public but feel personal. Aggregate before you display; never expose raw follower handles to clients in a UI.

Budget projection

For an agency tracking 200 creators across both platforms with a daily refresh:

• TikLiveAPI side. 200 creators x 1 /userinfo-by-username/ + 1 /user-posts/ page per day = 400 calls/day = ~12,000 calls/month. Plus weekly follower sampling at 5 cursor pages each = ~4,000 more. Roughly 16K credits/month. Pricing on /pricing/.
• Meta side. Instagram Graph API for Business is included with a verified Business app at no per-call fee, but rate limits apply (200 calls per user per hour). Plan around the limits, not around dollar cost.
• Hidden cost. The manual review queue. Budget 4 to 8 reviewer hours per week if you onboard new creators continuously.

Want the full endpoint catalog before you commit? Browse /documentation/ or read the rest of the /blog/ for adjacent patterns.

FAQ

Does TikLiveAPI call Instagram?

No. TikLiveAPI is TikTok-only. Every endpoint listed in the docs returns TikTok data. Instagram data must come from Meta's own APIs. This post is about combining the two cleanly, not about a single provider doing both.

Which Meta API should I use for Instagram?

For business and agency use, the Instagram Graph API for Business inside Meta Graph API is the right answer. The Instagram Basic Display API is creator-token-scoped and only useful for accounts whose owner authorizes you directly.

How accurate is the audience overlap estimate?

Directional. It is good enough to compare creators against each other and to prevent obvious double-counting in briefs. It is not good enough to publish as a hard percentage to a brand.

Can I match creators automatically with 100% confidence?

No. Bio-link cross-reference is the closest you get to certainty. Always keep a manual review path for ambiguous cases, and store who reviewed what.

What is the canonical TikTok identifier?

The numeric id returned by /userid/ and present inside user.id on /userinfo-by-username/. Store secUid as well, since some endpoints prefer it.

How do I authenticate TikLiveAPI requests?

Send your API key in the X-Api-Key header against https://api.tikliveapi.com. All endpoints are GET.

Where do I get help if my pipeline breaks?

Reach the team via /contact/ with the request URL, timestamp, and the masked first few characters of your API key.