All about Halo Effect

See TikTok content's lift on your Shopify/Amazon sales — how to connect each, and why Halo GMV differs from your dashboard.

The Short Answer

Halo Effect (Dashboard → Content → Halo Effect) puts your TikTok activity and your off-TikTok sales on the same daily timeline. You connect Amazon and/or Shopify, link your TikTok Shop products to their Amazon ASIN or Shopify variant equivalents, and Cruva overlays daily TikTok views and TikTok GMV against daily Amazon and Shopify revenue — plus a set of summary stats (correlation, estimated halo revenue, RPM, multiplier) that quantify how strongly they move together.

The one-sentence honest framing, which also answers "where is this number coming from?": Halo Effect is a correlation view, built on standard, transparent statistics — not a black box and not causal attribution. We don't (and can't) trace an individual Amazon purchase back to an individual TikTok video. What we can show is whether your off-platform sales rise and fall in step with your TikTok activity, and put honest numbers on that co-movement. Every formula is documented below.

Halo Effect is currently in beta and requires the Scale plan (shops that connected Amazon or Shopify before the gate keep access).

Why "Halo Effect"?

When creators post about your product on TikTok, not everyone buys in TikTok Shop. A large share of viewers search your brand on Amazon, or google you and land on your Shopify store. That spillover — demand created on TikTok but captured elsewhere — is the halo effect, and it's invisible in TikTok Shop analytics alone. If you only count TikTok GMV, you're undervaluing your creator program; this page exists to show the rest of the iceberg.

Setting Up

Step 1: Connect Amazon

From the Halo Effect page (or Dashboard → Integrations, or your shop's Integrations page), click Connect on the Amazon card.

You're sent to Amazon Seller Central to authorize Cruva via Amazon's official Selling Partner API consent screen. (Vendor Central / 1P accounts are supported too — the flow goes through Vendor Central instead.)

Approve, and you land back on the Halo Effect page connected.

Notes:

  • Sign in with the Amazon account that manages your Seller Central. If someone else runs your Amazon account, use "Copy Connection URL" on the Amazon card and send them the link — they can authorize without a Cruva login.
  • Cruva automatically maps your shop's region to the right Amazon marketplace (US, EU, etc.).
  • Reconnecting is safe; disconnecting is not. If your Amazon authorization expires (see troubleshooting), the Reconnect flow preserves all your product links. Clicking Disconnect, however, deletes the Amazon connection including its product links — you'd have to re-link products after reconnecting. If you're just fixing an auth error, always use Reconnect, never Disconnect.

Step 2: Connect Shopify

Shopify connects through the Cruva app on the Shopify App Store rather than a redirect flow. Two ways:

  • Connect it yourself: install the Cruva app from the Shopify App Store, then pick which Cruva shop to link it to.
  • Send a connect code: if the store owner is someone else, click "Generate connect code" on the Shopify card. This creates an 8-character code — single-use, valid for 7 days (generating a new one replaces it). The store owner installs the Cruva app and enters the code. No Cruva login needed on their side.

The first time you connect, Cruva pulls your recent Shopify order history in the background. You'll see a banner — "Pulling in Shopify data — we're pulling your last 30 days of Shopify orders for the first time. This usually takes a few minutes — the chart will update automatically." — and the page fills in on its own.

One-to-one rule: a Shopify store links to exactly one Cruva shop and vice versa. If you see "This Shopify store is already connected to another Cruva shop," disconnect it there first.

Step 3: Link your products and SKUs

This is the step people skip, and it's what makes the product-level view work. Go to My Shops → your shop → Integrations (the product filter's "Manage Linked Products" button takes you there too):

In the TikTok Products panel, search and select one of your TikTok Shop products.

In the Amazon panel, search the Amazon catalog (pre-filled with your product's name) and click Link on the matching listing (this stores the ASIN).

In the Shopify panel, search your store's products, expand one, pick the exact variant from the dropdown (you'll see the variant title, price, and SKU), and click Link Product.

How linking behaves:

  • One TikTok product can link to many ASINs and many Shopify variants. If your TikTok listing corresponds to three Amazon listings (sizes, bundles) or four Shopify variants, link them all — halo metrics aggregate revenue across every linked ASIN/variant, and the chart caption will say so (e.g. "Aggregating 3 Amazon products").
  • Duplicates are blocked ("This product link already exists") — only the exact same TikTok-product-to-ASIN (or to-variant) pair is rejected; additional different links are fine.
  • Linking a Shopify variant kicks off a background sync for that variant's sales history; give it a few minutes the first time.

You can use Halo Effect without linking anything — "All Products" mode shows whole-store Amazon/Shopify revenue against all your TikTok activity. Linking is what unlocks the per-product filter.

Reading the Page

The six stat cards

Every card has a ? tooltip explaining its math in-product; the full derivations are in the next section.

TikTok GMV — gross merchandise value generated inside TikTok Shop for the selected date range and product filter. This never includes Amazon/Shopify sales.

TikTok Views — total video views across your TikTok Shop affiliate videos (for the selected products) in the range.

Halo Correlation — how tightly daily TikTok views and daily off-platform revenue (Amazon + Shopify) move together, from 0% (no detectable relationship) to 100% (move in lockstep). Shows green at 30% or higher; shows "—" until there are at least 3 days of data.

Est. Halo Revenue — a statistical estimate of the off-platform dollars associated with your TikTok views over the period, from a regression of daily revenue on daily views. Shows "—" under 3 days of data; floored at $0.

Halo RPM — off-platform revenue per 1,000 TikTok views. A simple, comparable benchmark: "every thousand views coincided with $X of Amazon+Shopify sales."

Halo Multiplier — off-platform revenue per $1 of TikTok GMV. A 2.0× multiplier means for every dollar you see in TikTok Shop, two more dollars showed up on Amazon/Shopify in the same window.

The Revenue & Views chart

One daily timeline with up to four series: Amazon Revenue (orange), Shopify Revenue (green), TikTok GMV, and TikTok Views (indigos). Controls:

  • Normalize scales (on by default, under Customize): each line is rescaled to 0–100% of its own peak so you can compare shapes — a 50,000-view day and a $3,000-revenue day become visually comparable even though the units differ wildly. The hover tooltip always shows the real values. Turn it off to get true dollar/count axes (currency on the left, views on the right).
  • Show on chart: toggle individual series on/off (also click legend items). Your choices persist between visits.
  • Export: downloads the chart data as CSV for your own analysis.
  • Product filter: the dropdown next to the toolbar lists your linked TikTok products (searchable across TikTok, Amazon, and Shopify names). Selecting one filters TikTok stats to that product's videos AND filters Amazon/Shopify revenue to its linked ASINs/variants. If the product has no link on a platform, that platform's line shows $0 and the caption notes "No Amazon link" / "No Shopify link" — it does not fall back to whole-store revenue.

Click any point on the chart to open that day's video breakdown: a pie of the top 4 videos (plus an "Other" bucket) by GMV — or by views if the top videos had no GMV — with each video's thumbnail, creator handle, share of the day's total, view count, GMV, and a "View on TikTok" link. When a product is filtered you also get a direct "View on Amazon" / "View on Shopify" link. This is how you answer "what caused that spike on the 14th?" in two clicks.

The platform cards

Below the chart, each connected platform gets a card with its total revenue for the range (respecting the product filter) and a Manage Connection button. Unconnected platforms show a connect card instead.

Sharing

Halo Effect is included in Cruva's shareable dashboard: generate a share link and the recipient gets a read-only view of the stats, chart, and product-filtered data — no Cruva login, no dashboard access, and viewing a shared report never modifies your data.

Where the Numbers Come From (The Data Science)

This section exists because the most common question about this page is "where is this number coming from?" The answer: three well-defined daily data feeds, joined by date, run through statistics you'd find in any intro data-science course. Here's the whole pipeline.

The three data feeds

TikTok (daily): Cruva's analytics warehouse already tracks your shop's TikTok performance per day — total video views across your affiliate videos, and your TikTok Shop GMV (the true platform total for your shop; per-product totals when you filter). This is the same data behind the rest of your Cruva dashboard.

Amazon (daily): pulled from Amazon's official Selling Partner API order metrics — Amazon's own per-day totals of ordered revenue, order count, and units, for your whole marketplace or per linked ASIN. Two honest notes: (1) these are total Amazon sales, not "TikTok-attributed" sales — Amazon's Attribution program is not involved; (2) Amazon's own reporting lags roughly a day or two as orders settle, and Cruva caches Amazon responses for ~12 hours, so "today" and "yesterday" always look lower than they'll end up. Vendor Central (1P) accounts are special: Amazon provides no daily retail sales feed for vendors, so Cruva reconstructs a daily series from your purchase orders — each PO's value is spread evenly across its delivery window and smoothed with a 7-day moving average (totals are preserved exactly). Vendor lines therefore show shape and magnitude, not literal day-by-day consumer sales.

Shopify (daily): Cruva syncs your Shopify order history into a per-day summary (revenue, orders, units) — for the whole store, and separately for each linked variant. The sync runs nightly plus on-demand whenever the data is more than ~6 hours old (you get the slightly stale data instantly while it refreshes in the background), and the first-ever sync runs the moment you connect.

All three feeds are aligned by calendar day, giving one table: date | tiktok_views | tiktok_gmv | amazon_revenue | shopify_revenue. Every number on the page is computed from that table, over the date range you picked, and every stat updates when you change the date picker or the product filter.

"Active days"

Before computing the correlation and regression stats, days where both TikTok views and off-platform revenue are zero are dropped — a day with no signal on either side says nothing about the relationship. The remaining days are your sample size, shown in the Halo Correlation tooltip ("Based on N active days"). Correlation and regression need at least 3 active days; below that the cards show "—" rather than a meaningless number.

The formulas, exactly

Let each active day i have views_i (TikTok views) and rev_i (Amazon revenue + Shopify revenue that day).

Halo Correlation is the Pearson correlation coefficient (r) between the two daily series:

r = Σ (views_i − mean_views)(rev_i − mean_rev)

√[Σ (views_i − mean_views)²] × √[Σ (rev_i − mean_rev)²]

In plain English: for each day, we ask "were views above or below their average, and was revenue above or below its average?" If the two are consistently above-average together and below-average together, r approaches 1. If they're unrelated, r hovers near 0. The card shows r × 100%. Two deliberate product choices:

  • Floored at 0. A halo effect can't be negative by definition — TikTok presence shouldn't reduce your Amazon sales — so a negative r (which in practice means noise, seasonality, or an inventory event) is displayed as 0%, i.e. "no detectable halo," never as a negative contribution.
  • "—" when it can't be computed: fewer than 3 active days, or when one series is perfectly flat (constant views or constant revenue give the formula a zero denominator — you can't correlate against a flat line).

Est. Halo Revenue comes from an ordinary least-squares linear regression of daily revenue on daily views:

revenue = intercept + slope × views

The slope answers: "on average, each additional view coincided with how many additional off-platform dollars that same day?" The displayed value is:

Est. Halo Revenue = slope × (total views in the range)

— the revenue the model associates with your views, excluding the intercept (the intercept is your baseline daily revenue that happens regardless of views, so we deliberately don't credit it to TikTok). If the slope is zero or negative (more views didn't coincide with more revenue), the card shows $0. Like everything on this page, it's an association estimate, not an A/B-tested causal effect.

Halo RPM is pure arithmetic, no model:

RPM = (Amazon revenue + Shopify revenue) ÷ (TikTok views) × 1,000

Off-platform dollars per 1,000 views over the selected range. Shows "—" when views are zero. Because it's just a ratio of totals, it's the most robust of the halo stats — but also the least discriminating (it moves whenever either total moves, whether or not they're related day-to-day).

Halo Multiplier:

Multiplier = (Amazon revenue + Shopify revenue) ÷ (TikTok GMV)

How many off-platform dollars appeared per dollar of TikTok Shop GMV in the same window. Shows "—" when TikTok GMV is zero.

TikTok GMV / TikTok Views / Amazon Revenue / Shopify Revenue are straight sums of the daily values over the range (and product filter).

A worked example

Say over 14 active days you had 400,000 total views, $12,000 of TikTok GMV, and $18,000 of combined Amazon+Shopify revenue, and on days when views spiked, off-platform revenue reliably spiked with them:

  • Halo Correlation might read ~70% — strong co-movement.
  • The regression might find a slope of $0.03 per view → Est. Halo Revenue = 0.03 × 400,000 = $12,000 of the $18,000 is associated with view volume (the other ~$6,000 is the intercept — your baseline that we don't credit to TikTok).
  • Halo RPM = 18,000 ÷ 400,000 × 1,000 = $45 per 1,000 views.
  • Halo Multiplier = 18,000 ÷ 12,000 = 1.5× — every TikTok Shop dollar came with $1.50 off-platform.

Correlation is not causation — and why we say so

The tooltips repeat "correlational only — not causal" because it's true and it matters. This page has no per-purchase attribution (no pixel on Amazon, no coupon-code join), no control group, and no adjustment for seasonality, promotions, price changes, or your other ad channels. If you ran a Prime Day deal the same week your video went viral, both series spike and the correlation rises — the math can't tell those apart.

What makes the numbers useful despite that:

  • The same-timeline test is a real signal. Whole-store Amazon revenue that repeatedly jumps within a day of your TikTok spikes — across many spikes, on products you don't advertise elsewhere — is hard to explain any other way.
  • Product linking sharpens it. Filtering to one linked product turns "did my store move?" into "did this specific ASIN move when videos about this specific product got views?" — much closer to a natural experiment.
  • Conservative by construction. Negative correlations floor to 0, flat/negative slopes floor Est. Halo Revenue to $0, the regression intercept (baseline revenue) is never credited to TikTok, and nothing is shown at all below 3 days of data. When the page shows a halo, it's understating rather than inventing.

To strengthen your read: use longer date ranges (more active days = more reliable correlation/regression), compare a linked hero product against your store total, and click into spike days to confirm which videos drove them.

Freshness — "why doesn't this match my Seller Central right now?"

  • Amazon: metrics are cached ~12 hours on Cruva's side, and Amazon's order metrics themselves settle over ~1–3 days. Expect the most recent day or two to fill in over time.
  • Shopify: synced nightly, plus an automatic refresh whenever the snapshot is over ~6 hours old (stale data is served instantly while the refresh runs). Worst case you're looking at a few hours of lag.
  • TikTok: follows the same ingestion cadence as the rest of your Cruva dashboard, subject to TikTok's own reporting delay.
  • Currency & timezones: amounts are shown in each platform's native currency with no FX conversion (one more reason the normalized view is the default for cross-platform comparison), and day boundaries can differ slightly between your dashboard timezone and Amazon's UTC reporting — a sale at 11pm can land on the "wrong" side of midnight between systems.

So: small day-edge discrepancies vs. Seller Central or Shopify admin are expected and normal; totals converge once data settles.

Best Practices

Link products before you judge the feature. Whole-store mode is noisy — your Amazon total includes products TikTok never touches. The per-product view (one TikTok product ↔ its exact ASINs/variants) is where the halo becomes obvious or provably absent.

Link every ASIN and variant that maps to the TikTok product. Sizes, colors, bundles, multipacks — halo metrics aggregate across all links, so a partial mapping undercounts your halo.

Judge over 30+ days, not a week. The correlation and regression get sharper with more active days; 3 days is the mathematical minimum, not a recommendation. The 30-day default range is a good floor.

Use the stats as a set. High correlation + healthy slope = a real co-movement story. High RPM but ~0% correlation = you sell well off-platform, but not in step with your content — the "halo" may just be brand baseline. Multiplier is your executive-summary number ("every TikTok dollar brings $X.XX elsewhere"); correlation is your confidence check on it.

Investigate spikes with a click. Any interesting day on the chart → click it → see exactly which videos (and creators) dominated that day. That's your shortlist for who to retain, boost, or put on a campaign.

Compare periods, not just lines. Run the date picker over a big campaign window vs. a quiet month and compare RPM/multiplier between them — a before/after on your own numbers is often more persuasive than any single stat.

Keep the normalized view for shape, raw view for scale. Normalize on to see whether curves track; normalize off (or hover) when someone asks for actual dollars.

Hand connection links to the right people. Amazon's "Copy Connection URL" and Shopify's connect code exist so your marketplace managers can authorize without sharing logins. Codes are single-use and expire in 7 days.

Troubleshooting

Connection

"Your Amazon connection needs to be reconnected" (amber banner). Your Amazon SP-API authorization was revoked or expired (this happens when Seller Central permissions change or the authorization is removed on Amazon's side). Click Reconnect — it re-runs the Amazon authorization and keeps all your product links. Do not use Disconnect to fix this: disconnecting deletes the connection record along with its product links.

Amazon "Connection Error: Unable to sync with Amazon…" banner. Same family of problem — head to your shop's Integrations page and reconnect. If the error persists after reconnecting, contact support via the chat bubble.

"Failed to load Amazon metrics. Please check your SP-API credentials." The metrics call failed — usually auth (see above) or a transient Amazon API issue. Reconnect first, then retry.

"No pending install found. Open the app from your Shopify admin to finish linking." The Shopify-side install and the Cruva-side association got out of sync. Open the Cruva app from your Shopify admin — that re-mints the link — and finish the connection from there.

"Invalid or expired connect code" / "Connect code was just used." Codes are single-use and last 7 days, and generating a new one invalidates the old. Generate a fresh code and resend it.

"This Shopify store is already connected to another Cruva shop" (or the reverse). Shopify↔Cruva is strictly one-to-one. Disconnect the existing pairing first.

"Couldn't load Shopify data" toast. Usually transient — Cruva retries automatically. If it keeps happening, reconnect Shopify from Integrations.

Numbers

"Where is this number coming from?" See the data-science section above — every stat is a documented formula over three daily feeds (TikTok analytics, Amazon order metrics, Shopify order sync). Hover the ? on any stat card for its in-product explanation.

Halo Correlation / Est. Halo Revenue show "—". You have fewer than 3 active days in the selected range (days where views or off-platform revenue were nonzero), or one of the series is completely flat. Widen the date range.

Halo Correlation shows 0%. Either there's genuinely no day-to-day relationship in this window, or the underlying correlation was negative — which the page treats as "no detectable halo" rather than a negative effect. Try a longer range or filter to a linked product.

Amazon/Shopify shows $0 with a product selected. Check the caption under the chart title: "No Amazon link" / "No Shopify link" means the selected TikTok product has no mapping on that platform, and the filter shows $0 rather than silently falling back to store totals. Link the product (Manage Linked Products) and the revenue appears.

Recent days look too low vs. Seller Central / Shopify admin. Expected: Amazon order metrics settle over 1–3 days and are cached ~12h; Shopify syncs nightly/every ~6h. The last day or two always fills in late. Day-boundary (timezone) differences can also shift a late-night order to the adjacent day.

The chart is empty / "No data available for the selected date range." Check the date range (and product filter), and that at least one platform is connected. For a brand-new Shopify connection, wait for the "Pulling in Shopify data" banner to finish — the chart updates automatically.

Vendor (1P) daily revenue looks unnaturally smooth. By design: Amazon offers no daily retail sales feed for vendors, so your line is purchase-order value spread across delivery windows with 7-day smoothing. Totals are exact; day-to-day shape is modeled.

"This product link already exists." That exact TikTok-product ↔ ASIN/variant pair is already linked. To link the same TikTok product to a different ASIN or variant, just do so — multiple links are supported.

A newly linked Shopify variant shows no history. The first sync for a variant runs in the background right after linking; give it a few minutes.

Access

"Scale Plan Required" overlay. Halo Effect requires the Scale plan (shops with Amazon or Shopify connected before the gate are grandfathered in). Upgrade from Billing.

FAQ

Is this attribution? Can I say "TikTok caused $X on Amazon"? No — say "associated with." There's no purchase-level join to TikTok (Amazon Attribution isn't part of this). Est. Halo Revenue is a regression estimate of revenue that moves with your views; correlation quantifies co-movement. Strong, repeated, product-level co-movement is compelling evidence — but it is evidence, not proof.

Does "Amazon Revenue" include only TikTok-driven sales? No — it's your total Amazon ordered revenue (whole store, or for the linked ASINs when filtered). That's precisely why the correlation/regression stats exist: to estimate what share of that total moves with TikTok.

Can one TikTok product link to multiple ASINs or variants? Yes, as many as you like — metrics aggregate across all of them. Only exact duplicate pairs are blocked.

Can I edit a link? No — unlink and re-link.

What happens to my links if I disconnect? Shopify: connection is deactivated. Amazon: disconnecting deletes the connection and its product links — use Reconnect (not Disconnect) when fixing auth issues.

Do I need both platforms? No — either one lights up the page. Off-platform stats simply cover whichever is connected.

Can I share this with someone outside my team? Yes — via Cruva's shared dashboard link. It's read-only and doesn't require a Cruva account.

Can I get the raw data? Yes — the chart's Export button downloads the daily series as CSV.

Quick Reference

  • Plan requirement: Scale (existing Amazon/Shopify connections grandfathered)
  • Default date range: Last 30 days (timezone-aware picker)
  • Minimum data for correlation/regression: 3 active days (a day counts if views or off-platform revenue > 0)
  • Halo Correlation: Pearson r between daily views and daily Amazon+Shopify revenue, × 100%; floored at 0%; green ≥ 30%
  • Est. Halo Revenue: regression slope × total views (intercept excluded); floored at $0
  • Halo RPM: off-platform revenue ÷ views × 1,000
  • Halo Multiplier: off-platform revenue ÷ TikTok GMV
  • Amazon data: SP-API daily order metrics (total sales, not attribution); ~12h cache; settles over 1–3 days; vendors modeled from POs + 7-day smoothing
  • Shopify data: nightly sync + auto-refresh when > ~6h stale; ~1h cache; first sync runs on connect
  • Product links: one TikTok product → many ASINs/variants; aggregate; no edit (unlink + re-link); exact duplicates blocked
  • Shopify connect code: 8 characters, single-use, 7-day expiry, one live code per shop
  • Amazon disconnect: deletes connection and product links — use Reconnect for auth fixes
  • Chart: up to 4 series; normalize (0–100%) on by default; click a day for its top videos; CSV export
  • Currency: native per platform, no FX conversion

Was this article helpful?

Related articles