Is Self-Building Really Free? The 1-Year TCO of an EC Revenue Dashboard

"If we self-host Matomo, Umami, or GA4+Looker Studio we can build a free revenue dashboard, right? Why pay a monthly fee for RevenueScope?" We often hear this from Shopify EC operators in the 10M-50M yen GMV/month tier. The short answer: "OSS is free" is an accounting misconception — self-building generates roughly 460,000-800,000 yen of hidden 1-year cost (priced at the 5,000 yen/hr Japanese freelance marketer rate). This article lines up four options — Matomo, Umami, GA4+Looker Studio, and RevenueScope — on 1-year Total Cost of Ownership and maps the threshold at which an operator should switch.

TL;DR#

1. Self-hosting OSS or GA4+Looker Studio generates roughly 460,000-800,000 yen of 1-year TCO at the 5,000 yen/hr equivalent. RevenueScope Growth is 117,600 yen/year (9,800 yen × 12) — a 4-7× TCO advantage.
2. The hidden cost of "free" OSS is a 3-layer stack: opportunity cost on time (hours not spent on revenue activity), learning curve (Matomo configuration / GA4 event design / Looker calculated fields), and upgrade response (OSS major versions, GA4 API spec changes).
3. The decision doesn't end at TCO: operators who want to redirect engineer/operator hours to revenue activity → RevenueScope; teams that already have in-house engineers and SQL talent, or large enterprise / large-scale EC → self-build is rational. Most of this article's target readers fall into the former.

1. Why TCO comparison matters#

The "OSS is free, so cost is zero" view only counts software licensing. The real Total Cost of Ownership for an EC operator includes at least four elements beyond licensing:

• Initial setup hours: server configuration, tracking installation, dashboard build, QA
• Monthly operational hours: data quality checks, tracking fixes, new metrics, troubleshooting
• Server costs: VPS / cloud / storage
• Learning costs: documentation, troubleshooting research, internal knowledge transfer

Treated as work hours, "free" OSS still costs rate × time in human capital. At the 5,000 yen/hr freelance marketer / data analyst rate, self-building's annual TCO easily climbs into the hundreds-of-thousands-of-yen range [4].

Opportunity cost as a complement#

Beyond TCO sits opportunity cost. If an EC operator spends 40 hours on self-build, those hours can no longer go into ad creative, product-page improvement, or email marketing — all activities that move revenue. For an operator at 10M yen GMV/month, 40 hours is roughly 25% of one month's working days. We start from the recognition that "OSS license fee = 0" and "TCO = 0" are not the same thing.

2. The four options#

Four realistic options that Japanese SMB EC operators consider:

2.1 Matomo On-Premise#

Matomo is a GPLv3-licensed OSS Web Analytics tool, offered as an On-Premise edition you self-host and a paid Matomo Cloud edition [1]. This article evaluates the On-Premise TCO. Tracking, reports, and custom events are standard, but EC-focused Revenue / RPS (Revenue Per Session) / AOV dashboards are not in the standard template — you build them yourself with custom reports.

2.2 Umami v3#

Umami is an MIT-licensed lightweight Web Analytics tool [2]. It runs on Node.js + PostgreSQL, small enough to fit on free tiers of Vercel / Railway / Fly.io. Pageviews, bounce rate, and referrer reports are standard, but — like Matomo — there is no out-of-the-box EC Revenue dashboard, so custom events plus custom reports are required.

2.3 GA4 + Looker Studio#

GA4 is Google's free Web Analytics tool with native ecommerce event support. Looker Studio (formerly Data Studio) is a free BI tool that connects directly to GA4 for custom dashboards [3]. For deeper analysis, GA4 → BigQuery export with SQL aggregation is possible, though BigQuery query costs apply separately.

2.4 RevenueScope

RevenueScope is an EC-measurement SaaS designed for Japanese SMB EC. Out of the box it shows core 4 metrics — Revenue, AOV, RPS, CVR — plus Sessions in a 5-card KPI dashboard. The technical stack (GTM 5 minutes + dataLayer + GA4) is intentionally simple so a dedicated engineer is not required. (The "5 minutes" assumes GA4 ecommerce tracking is already set up.)

3. 1-Year TCO comparison table#

We line up the annual TCO of each option at the 5,000 yen/hr rate. Assumptions (industry-average estimates — actual hours vary by company; measure in your own environment):

• Hourly rate: 5,000 yen (median freelance contract rate). For internal staff time, read at roughly 10,000 yen/hr and double the TCO.
• Target environment: Tier-1 Shopify EC at 10M-50M yen GMV/month, running a revenue dashboard (channel-level Revenue / RPS / AOV / CVR).

* Industry-average estimates (not measured). Actuals vary with operational conditions, OSS version, engineer maturity, and existing assets (e.g., whether GA4 is already in production) — read this as an editable model.

3-1 Reading the numbers#

• Matomo Self-Host ≈ 800K yen: building EC-focused Revenue / RPS / AOV custom reports on a self-hosted On-Premise install. Matomo Cloud cuts setup hours but adds a license fee in the several-hundred-euro/month range.
• Umami / GA4+Looker ≈ 460-500K yen: lighter setup, but EC Revenue measurement still needs dataLayer integration, custom event design, and (for GA4+Looker) BigQuery and chart-design learning costs.
• RevenueScope Growth ≈ 120K yen: GTM 5-minute install (assumes GA4 ecommerce is already set up), existing dataLayer, zero extra configuration. Monthly ops is "look at the dashboard" — tens of minutes.

The intuition that "self-hosting OSS or GA4+Looker is free" collapses once you account for 40h of initial setup and 6-8h/month of operations.

4. When each option fits#

TCO alone is not the right judgement axis — each option has its own strengths. But the conditions under which DIY fits are narrower than they look. Most of this article's target readers (10M-50M yen GMV/month, 1-3 marketing operators, no dedicated engineer) will notice, reading below, that they do not qualify.

4-1 When Matomo On-Premise fits#

• Large enterprises where storing customer data on company-controlled servers is a hard requirement, and in-house engineers fluent in Linux server operation are on staff [5][6]
• Operators who want a self-designed revenue dashboard with custom metrics the standard offering does not cover
• Large-scale EC at 1B+ yen GMV/month, where self-hosting becomes more TCO-efficient than flat-rate SaaS

These three conditions holding simultaneously is, in practice, limited to large enterprises or large-scale EC.

4-2 When Umami fits#

• Lightweight Web Analytics is the goal and EC-specific Revenue measurement is not needed
• Solo developers / startups who want to run on free tiers of Vercel / Railway / Fly.io

If EC revenue judgement is the goal, Umami is out of scope from the start.

4-3 When GA4 + Looker Studio fits#

• Teams already invested in GA4 + BigQuery, with an analyst on staff who can write SQL
• Teams operating under a "stay within the Google ecosystem" constraint

Conversely, if that prerequisite is not yet in place, GA4+Looker becomes a "free-looking but learning- and operations-heavy" option (see the next section).

4-4 When RevenueScope fits

• Shopify / BASE / STORES / EC-CUBE operators in the 10M-50M yen GMV/month tier
• 1-3 marketing operators, no dedicated engineer
• GA4 ecommerce tracking already in place (zero additional configuration to adopt)
• Operators who want to make investment decisions on the core 4 metrics (Revenue / AOV / RPS / CVR) plus Sessions = 5 KPI cards
• Teams whose core operation is "look at the dashboard weekly and reallocate ad budget"

Putting this together, DIY pays off in one of two cases: you already have in-house engineers and SQL talent, or you are a large enterprise / large-scale EC. This article's target — 10M-50M yen GMV/month, no dedicated engineer — fits neither. To choose a tool on features rather than cost, How to choose an EC access-analytics tool is the companion piece. For RPS, see RPS (Revenue Per Session): formula and examples; for AOV, see What is AOV (Average Order Value).

5. Three traps in "free"#

Even when OSS or GA4+Looker Studio is "free on the books", real operations stack up costs in three layers.

5-1 Trap 1: Time opportunity cost#

If you spend 40 hours building Matomo, those hours leave revenue activity. The opportunity cost — A/B-testing ad creative, improving LPs, designing email segments — sits on a different axis from the "200,000 yen of initial setup" line and must be evaluated separately.

5-2 Trap 2: Learning curve#

Even with strong documentation, OSS and Looker Studio carry non-trivial learning costs the first time:

• Matomo: vast configuration surface; custom report syntax close to SQL
• GA4: event design, custom dimensions, dataLayer mechanics — all complex
• Looker Studio: calculated-field syntax; SQL knowledge once BigQuery is connected

Getting from "docs read" to "working dashboard" stacks up costs that never touch the license line.

5-3 Trap 3: Upgrade response#

OSS has periodic version bumps, GA4 has API spec changes, Looker Studio has UI revisions and connector spec changes:

• Matomo major version bump: schema changes can require migration work
• GA4 API spec change: BigQuery export schema shifts can break existing queries
• Looker Studio UI revision: chart configuration may need to be rebuilt

These are unplanned costs that occur a few times per year and are hard to fold into the "monthly ops hours" estimate. SaaS vendors absorb them, so user-side cost is zero.

6. Self-build vs SaaS decision flow#

We collapse the framing above into a flow you can run mechanically:

6-1 Q1: Are you a Shopify / BASE / STORES / EC-CUBE operator at 10M-50M yen GMV/month?#

• Yes → go to Q2
• No (under 10M yen GMV) → starting with GA4+Looker Studio is perfectly fine; the point where operations get heavy (the switching threshold below) is when to consider a move.
• No (1B+ yen GMV, large-scale EC) → consider BI tools (Tableau / Looker / Mode)

6-2 Q2: Do you want to redirect engineer / operator hours to revenue activity?#

• Yes → RevenueScope. 9,800 yen/month recovers roughly 460,000-800,000 yen of opportunity cost.
• No (you want to self-build OSS as a corporate philosophy) → Matomo / Umami / GA4+BigQuery. The moment you want those hours back for revenue activity is your review point.

6-3 Q3: Are core 4 metrics (Revenue / AOV / RPS / CVR) + Sessions = 5 KPI cards enough?#

• Yes → RevenueScope (5-metric specialization philosophy)
• No (you also want MMM / MTA / margin / LTV / inventory) → consider full-stack tools like Triple Whale (The 5 big trends in EC measurement for 2026 maps the full landscape)

Note: when ad spend is connected (ad API or manual input), RevenueScope shows per-channel ROAS, saturation, and a budget-allocation starting point on the same screen as your revenue metrics (no ROAS for periods without connected spend). ROAS here is revenue-based and limited to that connected-spend path; it does not reach MMM, MTA, or gross-margin-based ROAS — the design keeps operational load low by measuring in fewer places. See The Right Way to Design Marketing KPIs for details.

RevenueScope — where it sits: a substitute, not a complement

By this point, RevenueScope's positioning should be clear. It is not a complement to OSS or GA4+Looker Studio — it is a substitute built for Japanese SMB EC at 10M-50M yen GMV/month.

7-1 Why "substitute, not complement"#

• GA4 covers the traffic measurement / recording layer — that stays as GA4, and RevenueScope does not compete with it.
• The revenue dashboard layer built on top of Matomo / Umami / GA4+Looker Studio is what RevenueScope replaces. Same layer, same purpose (operating a revenue dashboard).
• Self-building generates 460,000-800,000 yen/year in TCO; RevenueScope handles the same operation for under 120,000 yen/year. Because the feature scope overlaps, the relationship is substitution, not complementation.

And RevenueScope returns more than the cost gap: the hours you save go back into revenue judgement, because per-channel RPS — and, for periods with connected ad spend, ROAS and a budget-allocation starting point — land on the same screen as your revenue. Handing over aggregation and the decision material is why it is not merely a "cheap substitute".

7-2 How to think about revenue-dashboard design#

Whether you build it yourself or use a SaaS, the metrics that belong on a revenue dashboard are the same. The decisive choice is what to show and what not to show — that determines whether the dashboard gets used long-term. RevenueScope deliberately narrows to "core 4 metrics + Sessions = 5 KPI cards". See The Right Way to Design a Revenue Dashboard for the full rationale.

7-3 Threshold for switching from self-build to RevenueScope

If you are already self-building, the value of switching can be judged on three points:

• You want to redirect 6h+/month of operational hours to revenue activity → switch recommended
• You spend 20h+/year on OSS / GA4 upgrade response → switch recommended
• The team has a "no one knows this configuration" knowledge silo → switch recommended

Conversely, if "OSS as a corporate philosophy" is non-negotiable, your engineers are already proficient, or your existing GA4 + BigQuery investment is large, continuing self-build is rational. Even so, most of this article's target readers (10M-50M yen GMV/month, no dedicated engineer) have already crossed at least one of the three thresholds above — and that is the signal to switch.

Summary#

• Self-building a revenue dashboard with OSS (Matomo / Umami) or GA4+Looker Studio generates roughly 460,000-800,000 yen of 1-year TCO at the 5,000 yen/hr equivalent, versus RevenueScope Growth at under 120,000 yen/year (9,800 yen/month) — 4-7× TCO efficiency
• The hidden cost of "free" OSS is a 3-layer stack: time opportunity cost, learning curve, upgrade response
• For 10M-50M yen GMV/month operators with no dedicated engineer who are satisfied with core 4 metrics, RevenueScope is the lowest-friction option; for corporate-philosophy OSS operators, mature in-house engineers, or 1B+ yen GMV/month large-scale EC, Matomo / Umami / GA4+BigQuery remains rational

The intuition that "OSS is free, so cost is zero" is a meaningful accounting misconception under TCO. For a 10M-50M yen GMV/month operator, 40 hours is a scarce resource that belongs on revenue activity. RevenueScope returns not just that time but the material — which channel, how much to allocate — for the revenue decisions you make with it, on the same screen. That is the essence beyond the TCO comparison.

For details on RevenueScope pricing, see the pricing page. For product philosophy, see About — the Revenue First view.

References#

[1] Matomo "Matomo On-Premise — Self-hosted Web Analytics" https://matomo.org/matomo-on-premise/

[2] Umami "Introduction" https://umami.is/docs

[3] Google "Data Studio documentation" https://cloud.google.com/looker/docs/studio

[4] Levtech Freelance "マーケティングのフリーランスの単価相場は？安定して案件を得る方法も紹介" (Japanese only) https://freelance.levtech.jp/guide/detail/31879/

[5] Ministry of Internal Affairs and Communications (Japan) "自分に関する情報が第三者に送信される場合、自身で確認できるようになります。" (Japanese only) https://www.soumu.go.jp/main_sosiki/joho_tsusin/d_syohi/gaibusoushin_kiritsu.html

[6] Personal Information Protection Commission (Japan) "個人情報の保護に関する法律についてのガイドライン（通則編）" (Japanese only) https://www.ppc.go.jp/personalinfo/legal/guidelines_tsusoku/