Looker Studio DTC templates worth building versus buying

Looker Studio (née Data Studio) is the default BI tool for warehouse-first DTC. It is free, it connects to BigQuery in one click, and it is the tool every incoming Google Ads consultant already knows. What it is not is infinitely scalable, and most DTC brands I work with end up in a build-versus-buy fight over the same handful of dashboards.

This is the decision log I walk through on every engagement. It answers the question "which templates should we build ourselves, which should we buy, and when should we graduate to Metabase or Hex?" It assumes a warehouse-first stack where BigQuery holds the source of truth.

The decision matrix

Six dashboards show up on every DTC operator's wishlist. For each, there is a clear build-versus-buy answer.

Let me walk each one.

Exec KPI overview: build it

Every DTC brand thinks it wants the same exec dashboard. Revenue, orders, AOV, conversion rate, CAC, LTV, ROAS. Five minutes in, the questions diverge. Revenue gross or net of refunds? AOV including shipping and tax or just merch? Conversion rate on unique visitors or sessions? Is returning-customer revenue in the CAC denominator?

There is no Looker Studio template that answers those questions in your brand's language. Every time I try to use one, the first half hour is renaming fields and the next two days are explaining to the team why the template's revenue number disagrees with the Shopify dashboard.

Build this from your own dbt mart models. Two or three pages max: one summary with the 8 to 12 KPIs that matter for your brand, one trend view with the same numbers over time, one cohort view for LTV.

The dbt mart models that feed this dashboard:

-- mart_exec_kpis.sql
SELECT
  DATE_TRUNC(created_at, WEEK) AS week,
  COUNT(DISTINCT order_id) AS orders,
  COUNT(DISTINCT customer_id) AS customers,
  SUM(total_price) AS revenue_gross,
  SUM(total_price - total_discounts - total_refunds) AS revenue_net,
  SUM(total_price) / COUNT(DISTINCT order_id) AS aov,
  SUM(total_price) / COUNT(DISTINCT customer_id) AS arpu,
FROM {{ ref('stg_shopify__orders') }}
LEFT JOIN {{ ref('int_refunds_by_order') }} USING (order_id)
WHERE financial_status = 'paid'
GROUP BY 1

Looker Studio reads this view directly. No joins, no calculated fields in the tool itself. If the exec asks "why is the number different from X," you have a single place to check and fix.

Product analytics: build it

SKU structure is too brand-specific for any template to fit. A brand with 20 SKUs across 3 categories needs a different dashboard than a brand with 2,000 SKUs across 40 categories. Template dashboards end up showing either too few products to be useful or too many to be readable.

The version I ship covers: SKU-level revenue and units, run rate, inventory days-on-hand, first-purchase rate (the percent of customers whose first ever purchase includes this SKU), and repeat-purchase lift (does buying this SKU correlate with higher LTV?).

The repeat-purchase lift query is the one that usually justifies the dashboard on its own. Run it in SQL first, see which SKUs cluster high-LTV customers, share it with the marketing team, watch them reshuffle the promo calendar.

Customer segmentation: build it

Cohort LTV, RFM segmentation, acquisition channel analysis. These live in your warehouse and the cohort LTV patterns article walks the SQL. Looker Studio's job is to render; the segmentation logic belongs in dbt.

The trap here is Looker Studio's "Calculated Fields" feature, which lets you define segments in the dashboard itself. Do not use it for segmentation. Every calculated field in Looker Studio is opaque to downstream tools and breaks when you share the dashboard. Put the segmentation in a dbt model, expose the segment as a column, use Looker Studio's filter controls.

Marketing funnel: buy, customize

Google's own Google Ads + GA4 Looker Studio template is genuinely good. It covers the default funnel (impressions → clicks → sessions → add-to-carts → purchases), per-campaign ROAS, and a reasonable attribution comparison view. For 80 percent of DTC brands, this template plus 20 minutes of customization is enough.

The customization needed: swap the default GA4 data source for your own warehouse view (not GA4 direct), add Meta conversion data from your warehouse, filter to your brand's actual campaigns. If you are already reading Meta, TikTok, and Google ads from your warehouse via dbt staging models, this is a 30-minute exercise.

Geographic performance: buy, leave alone

The GA4 geographic template is fine as-is. City-level heatmaps, country-level revenue tables, region-over-region growth. Nothing custom to DTC. The only customization worth doing is filtering to the geos where you actually ship.

Real-time monitor: graduate

Looker Studio is not a real-time tool. Its query cache is aggressive (default 12 hours) and refresh is manual or on a schedule. For a dashboard that shows live revenue as orders come in, you need something that streams.

Options that work: Metabase with live database connections, Hex with Python notebooks wired to streaming data, Grafana with BigQuery BI Engine reserved, or a custom Next.js page reading from BigQuery Storage API. I usually ship this as a custom page rather than a BI tool because the queries are simple and the team cares about polish.

If you need a real-time monitor on the cheap, a static Next.js page that polls a single BigQuery view every 60 seconds is $5 a month and handles most real-time DTC needs. The analytics engine case study has a version of this running in production.

When to graduate off Looker Studio entirely

Looker Studio starts breaking at three specific thresholds.

50+ dashboards. Looker Studio has no good project-level organization. Past 50 dashboards, finding the one you want becomes painful. Metabase's collections and Hex's folders handle this better.

Multi-tenancy for clients. If you are a fractional operator running dashboards for multiple brands, Looker Studio's sharing model is brutal. Every dashboard has to be individually shared and permissioned. Metabase with its SSO and row-level security is a better fit here. Hex with its workspace model is another.

Python in dashboards. Any dashboard that needs a statistical model (cohort projection, Bayesian CLV, etc.) is easier to build in Hex than to force into Looker Studio's charting primitives. Hex is the one I graduate to first; $24 per developer per month as of April 2026, and the notebook-plus-charts model fits an analytical workflow.

FAQ

What to try this week

Pick one of your existing Looker Studio dashboards. Open the Data tab. Check whether it reads from a raw table or a view. If it is raw, create a dbt staging view that matches what the dashboard actually uses, point the dashboard at the view, and watch the query cost drop.

If you do not have a warehouse yet, and you are running Looker Studio directly against GA4 and Shopify connectors, that is the deeper problem. A DTC Stack Audit will tell you whether your dashboards are the symptom or the cause.

Sources and specifics

• Looker Studio pricing: free tier (data-studio.google.com), Pro at $9 per user per month as of April 2026.
• Hex pricing: $24 per developer per month as of April 2026 (hex.tech).
• Metabase: open-source free, Cloud from $85/month.
• Template decisions above are from multiple DTC analytics rebuilds; the three-build-three-buy split holds across mid-market brands but varies for enterprise and headless.
• Real-time monitor pattern is from the Q1 2026 analytics engine case study.