Glamnetic · TikTok Shop · Omni

Graph ideas for the TikTok leadership dashboard

Eleven visualizations built from the fields already sitting in your Omni model — each one checked against what Omni can actually render, and each one mocked up against real July 2026 data so you can see it before you build it.

For Ellie Hu · TikTok Growth Strategy From Steph Pierson · Kombo Ventures Source dash omniapp.co/dashboards/8a31eda0 Data window July 1–31, 2026
01 / METHOD

How this was put together

The ideas below aren't the interesting part — the process is. This is the sequence I use whenever I need options fast on something I don't have deep domain knowledge in. It works precisely because it doesn't rely on me already knowing the answer.

1

Start from the artifact, never a blank page

First move was opening your actual dashboard and listing every saved query behind it. Nine of them. Brainstorming in the abstract produces charts that die on contact with the data model.

2

Pull the real numbers first

I ran your dashboard's own queries plus a few fresh ones straight out of Omni. Every mockup on this page is drawn with genuine July data — so the shapes are real, and a bad idea reveals itself as a flat, boring chart immediately.

3

Fence in what's actually buildable

Before generating a single idea, I had the tool go read Omni's documentation and come back with a verified list of chart types and options. An idea you can't ship is worse than no idea — it costs a meeting to discover.

4

Generate wide, then cut hard

Ask for more than you need, with the question each chart answers attached. The question is the filter. Anything that doesn't change a decision gets cut, no matter how good it looks.

5

Make it reviewable, not describable

Last step is rendering it. A described chart gets a polite nod; a drawn chart gets a real reaction. This page exists so you can point at things and say "that one, not that one."

The part worth stealing: steps 2 and 3 are what separate this from a generic listicle of "10 great TikTok charts." Grounding in your real fields means every idea is buildable, and constraining to Omni's real capabilities means none of them waste your afternoon.

Everything after that is just judgment — which is the part you keep for yourself.

02 / BASELINE

What the dashboard already covers

Your workbook holds nine saved queries. They are genuinely comprehensive on measurement — the gap is that almost everything is a daily grid, which reads as a spreadsheet rather than an argument. Leadership reviews need charts that carry a conclusion.

The nine existing queries

Dashboard 8a31eda0 · workbook contents
  • daily report27 measures · daily grid
  • daily report Copyduplicate
  • samplesseeding volume
  • contentposts by day
  • content by affiliatecreators, LIVEs, videos
  • affiliate revenuescreator GMV
  • shopify reportingTTS gross / net rev
  • TT GMV Max Spenddaily ad spend
  • Impressions / PDP Performancefull funnel

What's missing for a leadership review

The gaps these eleven ideas are aimed at
  • Nothing puts spend and return on the same axis — efficiency is invisible
  • The funnel data exists but is never drawn as a funnel
  • No view shows creator concentration, so roster risk is unmeasured
  • Creator productivity is aggregated to a daily total, hiding the outliers
  • No period-over-period framing — every number is an absolute with no verdict
  • Product mix lives in the model but has no tile at all
03 / THE IDEAS

Eleven charts, every one buildable in Omni

Each card carries the question it answers, the exact Omni chart type and fields, a mockup drawn on real July data, and the finding that mockup surfaced. The green badge means I verified the chart type and every option it needs against Omni's documentation.

01

GMV vs. GMV Max spend, with ROAS on a second axis

Are we buying this growth or earning it?

Bar + line combo Native · verified
Mockup · real July 2026 data
Total GMV GMV Max spend ROAS (right axis)

Omni build

Bar chart, then in Series configuration set ROAS's mark type to line and assign it to the right Y-axis. Omni lets each series pick its own mark type and its own axis.

Fields

  • X · Date (day)
  • Y left · Total GMV
  • Y left · $ TikTok Shop GMV Max Spend
  • Y right · Shop ROAS

What it surfaced

Daily spend climbed 77% across July ($5.9K → $10.4K) while GMV climbed 103% ($9.5K → $19.3K). Efficiency improved rather than eroded — that's the headline, and right now no tile says it.
02

KPI header strip with period-over-period and sparklines

In five seconds, is the month good or bad?

KPI tiles Native · verified
Mockup · July vs. prior period

Omni build

Omni's KPI tile is a component builder, not just a big number. One tile can stack a value, a comparison delta, and a sparkline chart. Five of these across the top reframes the whole dashboard.

Fields

  • Number · Total GMV · Orders · AOV
  • Number · Shop ROAS · Aff // Refund %
  • Comparison · prior period
  • Chart · sparkline, same measure

What it surfaced

This is the single highest leverage change on the page. It costs about twenty minutes and it's the difference between leadership reading numbers and leadership reaching a verdict.
Watch out: KPI dynamic number sizing isn't supported on Omni's classic dashboards — set sizes explicitly.
03

The TikTok Shop conversion funnel

Where exactly does TikTok attention leak away?

Funnel Native · verified
Mockup · July 2026 totals

Omni build

Funnel is a first-class visualization type in Omni — no custom work needed. Your Impressions / PDP Performance query already holds all three stages; it's just being rendered as a grid.

Fields

  • Stage 1 · # Impressions
  • Stage 2 · # Product Page Views
  • Stage 3 · # Items Sold

What it surfaced

21.3M impressions produce 831.5K product page views — a 3.9% click-through. Of those, 3.76% convert to a sale. The drop from impression to PDP is where the money is: a single point of CTR is worth roughly 8,000 extra page views a day.
04

Creator concentration (Pareto)

How exposed are we if our top creators walk?

Bar + cumulative line Native · verified
Mockup · top 10 creators, July 2026
Creator GMV Cumulative % of all creator GMV

Omni build

Bar chart sorted descending, plus a running-total table calculation expressed as a percentage, set to a line mark on the right axis. Both the combo mark type and the second axis are standard Omni configuration.

Fields

  • X · Creator
  • Y left · $ Creator GMV
  • Y right · running total % (table calc)

What it surfaced

Ten creators out of 3,121 drive 38.6% of all creator GMV — $110.8K of $287.2K. That's a roster risk nobody is currently looking at, and it's the kind of number that changes how a retention budget gets argued for.
05

GMV mix by content source, stacked

Is growth coming from LIVE, from video, or from the shop tab?

Stacked area Native · verified
Mockup · shape is illustrative, see note
Video GMV LIVE GMV Linkshare GMV Shop GMV

Omni build

Area chart with stacking set to Stack. Switch the same tile to Stack % to read mix instead of volume — that second view is often the more useful one for a review.

Fields

  • X · Date (day)
  • Y · Aff // Video GMV
  • Y · Aff // Live GMV
  • Y · Aff // Linkshare GMV
  • Y · Aff // Shop GMV

Why it matters

All four fields already exist in your daily report query. This tile is the one that tells you where to point creator resource next quarter, and it's currently four columns of a spreadsheet nobody reads across.
Honest caveat: these four measures are calculated fields that only resolve inside the Omni workbook, so the mockup above uses an illustrative shape, not real values. Every other chart on this page uses real data. Also: in Stack % mode Omni's tooltips show raw values, not percentages — add a "% of row" table calc if that matters.
06

Content volume vs. GMV per post

Are we getting more from creators, or just more out of them?

Stacked bar + line Native · verified
Mockup · real July 2026 data
Videos posted LIVE streams GMV per post (right axis)

Omni build

Two measures as stacked bars, a third as a line on the right axis. Stacking mode Stack, series mark type line for the ratio.

Fields

  • X · Date (day)
  • Y left · # Videos · # LIVE Streams
  • Y right · GMV Per Content Post

What it surfaced

This is the most surprising chart of the eleven. Content volume fell 22% across July (368 posts/day → 286) while GMV per post tripled, from roughly $16 to $47. The roster got smaller and dramatically better. Nothing on the current dashboard would let you notice that.
07

Day-of-week heatmap

When should we be spending, seeding and going live?

Heatmap Native · verified
Mockup · GMV by weekday × week, real July data

Omni build

Heatmap is a native Omni type. Day-of-week on one axis, week on the other, GMV as the colour measure. Use the colour scale tab to pin the midpoint at the daily average so above/below reads instantly.

Fields

  • X · Date (day of week)
  • Y · Date (week)
  • Colour · Total GMV

What it surfaced

The week has a hard rhythm. Thursday–Sunday carry the month; Sunday 19 July did $19.9K against Tuesday 21 July's $7.7K — a 2.6× swing on the same product, same week. That's a spend-pacing and creator-scheduling decision hiding in plain sight.
08

Top products with in-cell bars

What is TikTok actually selling for us?

Table · in-cell bars Native · verified
Mockup · real July 2026 data

Omni build

A Table visualization with the numeric column's display set to bar. Omni renders in-cell bars natively and auto-scales them. Add conditional formatting on the units column for a second read.

Fields

  • Row · Product Name
  • Measure · $ Sale Price (bar display)
  • Measure · # Units

What it surfaced

Short Oval leads at $45.2K, but the real signal is Fright Night at $37.7K in July — Halloween demand is landing three months early on TikTok. If merchandising is planning to that curve, this tile is how they'd know.
09

Refund rate against a target line

Is quality holding as volume scales?

Line + reference line Native · limited config
Mockup · real July 2026 data
Daily refund rate July average (5.23%)

Omni build

Line chart with a reference line configured on the Y-axis. Omni supports reference lines on both axes.

Fields

  • X · Date (day)
  • Y · Refund Rate
  • Ref line · target threshold

What it surfaced

Refund rate swings violently — 1.5% to 12.5% inside one month, against a 5.23% average. That volatility is itself the finding; a monthly average would hide it completely.
Watch out: Omni's own docs note reference-line configuration is limited. For a proper target band rather than a single line, add a table calculation as an extra series.
10

Creator efficiency scatter

Which creators are efficient, and which are just loud?

Scatterplot · bubble Native · verified
Mockup · top 10 creators, real July data · bubble = units sold

Omni build

Scatterplot with bubble sizing — a native type. Videos on X, GMV on Y, items sold driving the bubble size.

Fields

  • X · # Videos
  • Y · $ Creator GMV
  • Size · # Items Sold
  • Point · Creator

What it surfaced

The spread is enormous. sunflowerhigh did $11.6K on a single video. sparklesteph77 did $7.1K on 116 videos. Comparable GMV, 116× the content. Averages bury this; a scatter makes it the first thing you see — and it's a direct input to how creator deals get structured.
11

Small multiples by collection

Which collections are trending up, side by side?

Trellis / small multiples Native · undocumented
Mockup · GMV trend faceted by collection

Omni build

Omni has a native Small Multiples section in chart configuration, with independent scales and its own labels tab. Set the facet field to Product Name and it splits one query into a grid of mini-charts.

Fields

  • X · Date (day)
  • Y · $ Sale Price
  • Facet · Product Name

Why it matters

Eight collections on one line chart is spaghetti. Eight small charts is a briefing. This is the tile that replaces "how's the portfolio doing" with an answer.
Flagging honestly: small multiples is confirmed shipped and generally available via Omni's changelog, but it has no documentation page, so I couldn't verify which chart types support it. Worth a two-minute test before you promise it to anyone. Vega-Lite has a documented trellis example as a fallback.
04 / REFERENCE

What Omni can and can't render

Pulled from Omni's own documentation so nothing above is wishful thinking. Useful beyond this project — it's the constraint list for any future tile you're asked for.

CapabilityStatusNotes
Core chart typesNativeBar, line, area, pie & donut, scatterplot & bubble, boxplot, funnel, sankey, heatmap, KPI, table, single record, maps, markdown tile, AI summary — fifteen types, each with its own documentation page.
Pivot tablesNativeNot a separate type — it's the Table viz with fields dragged into a Pivots section.
Combo (bar + line)NativeNot a chart type either. Each series independently picks bar, line, scatter or area.
Dual / secondary axisNativeSeries are assignable to a left or right Y-axis. Powers ideas 01, 04 and 06.
Stacking modesNativeAutomatic, Stack, Group, Overlay and Stack % (100%). Group is bar-only.
Reference linesLimitedAvailable on both axes, but Omni's docs call the configuration limited and suggest a table calculation for anything complex.
Trellis / small multiplesShipped, undocumentedGenerally available per the changelog, with sorting and independent scales — but no doc page exists, so coverage across chart types is unverified.
TrendlinesNativeMoving average and regression under a per-series Analytics tab. Note: R² and the formula can't be displayed.
Conditional formattingNativeRow/column rules plus gradient colour scales with pinned min, mid and max.
In-cell bars in tablesNativeNumeric columns render as bars, auto-scaled or fixed. Powers idea 08.
SparklinesKPI onlyAvailable as a KPI chart component. Not a table column option — a sparkline inside a table needs a Markdown viz.
Period-over-periodSupportedVia KPI comparison components and dashboard controls. No dedicated doc page — don't point anyone at one.
Waterfall · gantt · gauge · bullet · sunburst · word cloudVega-LiteNot native, but all have worked example specs in Omni's custom-visualization gallery.
HistogramVia binningNo histogram type and no Vega example. Achievable as a bar chart over a binned dimension.
TreemapUnverifiedNo type page, no changelog entry, no Vega example. A demo title mentions it. I would not promise one.
Time-series bar chartsGotchaBar X-axes are ordinal, so missing dates simply don't plot. Fill missing data explicitly or use a line.
Custom Vega-Lite editorTwo costsOpening the advanced editor breaks the existing visualization (copy the query first) and disables drill-down. Specs don't auto-save.
05 / SEQUENCE

What I'd actually build first

Eleven is more than anyone should ship at once. If it were my dashboard I'd stage it like this — earning the right to the harder tiles by making the easy ones land.

Phase one · this week

Make it read like a verdict

  1. KPI header strip (idea 02)
  2. GMV vs. spend with ROAS (idea 01)
  3. Conversion funnel (idea 03)

All three are native, fast, and use fields already in the workbook. Together they turn a measurement grid into something a leadership audience can react to in the first thirty seconds.

Phase two · next

Answer the creator questions

  1. Content volume vs. GMV per post (idea 06)
  2. Creator efficiency scatter (idea 10)
  3. Creator concentration Pareto (idea 04)

These three carry the most genuinely new information — the July data already shows a roster that got smaller and much more productive, plus real concentration risk. This is the phase that earns the dashboard a standing slot.

Phase three · once it's landed

Fill in the texture

  1. Top products table (idea 08)
  2. Day-of-week heatmap (idea 07)
  3. Refund rate vs. target (idea 09)
  4. GMV mix by source (idea 05)
  5. Small multiples (idea 11)

Valuable, but each answers a narrower question. Build them when someone asks — a dashboard earns tiles, it shouldn't start with them.