RoAS vs ROI which metric should a business owner track

«For five out of ten advertising budgets, an increase in spending does not lead to an increase in profit». This finding is often confirmed in our projects at BUSINESS SITE and echoes foreign studies on diminishing returns: channels become saturated, and metrics plateau. At the same time, the “advertising works” report looks convincing: RoAS above 3, revenue is growing, reports from ad accounts shine. The business owner sees positive dynamics and asks a reasonable question: why is the balance in the PrivatBank account still not so rosy?

When a manager sets the single KPI “raise RoAS”, the team often improves the visible metric and loses profit due to discounts, logistics overspend, and cannibalization of organic traffic. When the focus is only ROI, marketing loses responsiveness, and creatives “burn out” before they are refreshed.

The purpose of this guide: to help an owner, CEO, or marketing director choose the metric to solve specific tasks and assemble a unified KPI system. I will show formulas, thresholds, examples from BUSINESS SITE practice, and ready-made templates for Excel/BI and SQL/ETL. As a result, you’ll get a clear logic: when to scale by RoAS, when by ROI, how to account for LTV and CAC, where to apply incrementality and MMM, and how to design an executive dashboard so that the morning decision matches the evening P&L.

ROAS for advertising: definition and formula

I often use RoAS as an operational indicator to scale up what is already «working» within a week and reduce budgets for unprofitable groups.

Blend‑ROAS (aggregated across all channels) is useful for weekly monitoring: it shows the total return from advertising. But management decisions are better made by breakdown: ROAS by channels, campaigns and even by SKU/categories. In our projects blend‑ROAS often looks stable while inside there is overheated branded search and weak contribution from social networks. Breakdown saves the budget from illusions.

Margin, RoAS, and COGS adjustment

To convert gross RoAS to margin‑adjusted RoAS, it’s useful to proceed step by step:

1. Gross Profit = Revenue − COGS − Discounts.
2. Contribution Margin = Gross Profit − Logistics − Average cost of returns.
3. mROAS = Contribution Margin / Advertising spend.

Mini‑example (e‑commerce):

• Revenue: 1,000,000 UAH
• COGS: 600,000 UAH
• Discounts: 50,000 UAH
• Logistics (including Nova Poshta): 80,000 UAH
• Returns (product + return shipping): 40,000 UAH
• Advertising spend: 150,000 UAH

Calculation:

• Gross Profit = 1,000,000 − 600,000 − 50,000 = 350,000
• Contribution Margin = 350,000 − 80,000 − 40,000 = 230,000
• gross RoAS = 1,000,000 / 150,000 = 6.67
• mROAS = 230,000 / 150,000 = 1.53

ROAS by channel and blended ROAS

Segmentation by channels — search, social networks, contextual media, marketplaces, affiliates — shows where advertising creates additional revenue and where it redistributes organic orders. For example, increasing bids on branded search for the store name in Google can easily improve blend‑ROAS and worsen true efficiency: some customers would have come anyway via a direct visit or through Rozetka.

In BUSINESS SITE projects we often observe cannibalization: remarketing «takes the cream off» users who already came from email or organic. The solution: combined methods — weighted blend (weighting by share of assists) and a contribution attribution model where the top of the funnel (video, reach) is assigned a share of influence.

ROI: definition and meaning

How to calculate ROI for advertising and marketing

For marketing it’s important to separate direct and indirect effects, as well as time horizons:

• Short‑term campaigns (promotions, sales): account for quick sales, returns and logistics, attribution window 7–14 days.
• Long‑term investments (branding, content, seo): plan horizons of 3–12 months, use ROMI/ROI, MMM or incrementality.

What data is needed for a correct ROI calculation:

• Revenue by channels and campaigns (by orders/MRR/ARR for subscriptions).
• Cost of goods sold (COGS), discounts, fulfillment, returns, marketplace fees.
• OPEX: salaries, software, agency fees, warehouse, courier expenses, payment processing fees (for example, Monobank/PrivatBank).
• Cost of capital (WACC) for NPV/IRR.
• Single source of truth: data warehouse, ETL from advertising APIs, crm and payment systems.

ROI from discounts, returns and logistics

Formula for adjusted ROI accounting for cost of capital:

• Free Cash Flow from the initiative by periods → discounting by WACC → NPV.
• ROI_adj = NPV / Investments.
• To compare initiatives we also apply IRR and ROIC to assess the alternative cost of capital.

RoAS vs ROI: differences, pros and cons

Cases when RoAS can mislead the owner:

• Overuse of branded search and remarketing, cannibalization of organic traffic.
• High gross RoAS with low margins and expensive logistics.
• Short attribution window with a long deal cycle.

Cases when ROI isn’t sensitive enough:

• A/B tests of creatives where it’s important to see the effect daily.
• Quick promotions and sales requiring rapid budget reallocation.
• Channels with strong seasonality where decisions are better made by mROAS.

Examples of conflicting signals between RoAS and ROI

Checklist when signals conflict:

• Check margin and COGS, calculate mROAS.
• Assess LTV and payback, account for retention rate and churn.
• Look at cannibalization and the contribution of the top of the funnel.
• Decide based on unit economics: scale, optimize, or run an incrementality test.

When a business should focus on ROAS

RoAS is the main KPI for performance marketing with a fast cycle: promotions, remarketing, DSA, catalog campaigns, A/B tests of creatives in social media. It helps manage bid strategies and automation in ad systems where algorithms optimize for a target ROAS.

It’s appropriate to adjust RoAS benchmarks by margin across verticals: for fashion with a margin of 50–60% the gross RoAS threshold can be lower than for electronics with 10–20%.

A quick decision rule: if mROAS is above the threshold and holds when increasing budget by 10–20%, scale stepwise and monitor the frequency cap and creative fatigue. If mROAS falls as the budget grows, the algorithm needs fresh creative, a new audience, or a bid adjustment.

When should a business owner look at ROI?

Examples of KPIs and solutions:

• Brand GRP/Reach + ROMI for the quarter and the impact on organic traffic and direct traffic.
• ROI for launching a new category taking into account COGS, logistics and marketplace fees.
• Capital allocation: distribution of budget between channels taking into account opportunity cost and warehouse constraints.

LTV, CAC and LTV:CAC: choosing RoAS or ROI

LTV (lifetime value of a customer) and CAC (customer acquisition cost) shift the focus from a one-time purchase to the lifecycle. For subscriptions and repeat purchases LTV:CAC determines the sustainability of growth. Benchmark thresholds: for e‑commerce LTV:CAC ≥ 3 is considered sustainable, for SaaS 3–5 with payback ≤ 12 months (6–9 is better for SMB).

Formula LTV:CAC:

• LTV = ARPU × gross margin × average retention duration (or 1/churn).
• LTV:CAC = LTV / CAC.
• Payback period = CAC / (ARPU × gross margin − variable servicing costs).

Impact of privacy changes on RoAS/ROI

In browsers, cookie deprecation creates a gap between clicks and purchases, especially in cross‑device.

Alternatives are server‑side tracking and conversion APIs, which improve data quality, as well as conversion modelling and data‑driven attribution to make up for signal loss. I recommend combining models: operationally look at last‑click/micro‑attribution, strategically – multi‑touch and MMM, to see upper‑funnel contribution and adjust RoAS/ROI thresholds.

Impact of attribution methods on RoAS

• Last‑click: simplicity and transparency, but it overvalues the lower funnel. Useful for operational control.
• Multi‑touch/data‑driven: distributes credit between touches, accounts for assists; more accurate, but requires high‑quality data.
• MMM (marketing mix modeling): an aggregated model at the channel/region/week level, accounts for seasonality and external factors, suitable for strategy and budgeting.

The choice of model depends on data maturity and funnel structure. For e‑commerce with significant offline contribution or omnichannel fulfillment MMM provides stable ROI benchmarks, and multi‑touch helps allocate budgets fairly between channels.

Incrementality assessment: Incrementality and iROAS

Design: holdout‑groups and conversion lift studies, where part of the audience is excluded from impressions or receives pseudo-impressions. For statistical power, we calculate sample size, choose a p‑value (usually 0.05) and the experiment duration taking seasonality and conversion lag into account.

Step-by-step plan for implementing incrementality‑testing:

1. Choose a channel/campaign with a significant share of the budget.
2. Form groups: test and control (holdout), ensure randomization.
3. Define the effect metric: orders, revenue, MRR/ARR, LTV.
4. Launch the test, collect data, check statistical power.
5. Calculate iROAS and update thresholds for scaling or cutting budget.
6. Enter results into MMM/attribution to reconcile models.

iROAS calculation: example and interpretation

Suppose a social media campaign: budget 300 000 UAH. The test group produced 1 200 orders with revenue 1 800 000 UAH, the control group, 900 orders with revenue 1 350 000 UAH. Incremental Revenue = 1 800 000 − 1 350 000 = 450 000 UAH.

How to adjust the budget: increase share for segments with the highest uplift, reduce it for groups with low responsiveness. It’s convenient to combine test results with attribution models: iROAS sets the “ceiling” of a channel’s contribution, while multi‑touch distributes contribution among touches.

Budget optimization by ROAS and ROI

Sensitivity analysis helps test the robustness of the decision: what will happen to ROI if CPC changes by 15% or returns increase by 5 percentage points. In BUSINESS SITE we add automation rules: if a channel’s mROAS falls by X% for three consecutive days: lower bids; if iROAS confirms uplift: allow budget growth by Y% per week.

Инструменты: bid strategies в рекламных системах, правила в google ads/Meta, predictive analytics в BI для прогнозирования дохода и нагрузки на операционный блок. Такой стек экономит время и стабилизирует KPI.

When to Scale Advertising: ROAS or ROI

Критерии для RoAS‑масштабирования:

• Short purchase cycle and confirmed mROAS above the threshold.
• Low marginal cost of acquisition (marginal CAC) when increasing budget.
• Sufficient operational capacity (warehouse, logistics, support).

Критерии для ROI‑масштабирования:

• Long LTV horizon, subscription model or frequent repeat purchases.
• LTV:CAC ≥ 3 and payback ≤ 9–12 months.
• Confirmed uplift via incrementality/MMM.

Examples: for e‑commerce: scale by RoAS adjusted for margin and seasonality; for SaaS: prioritize ROI with a focus on LTV:CAC, retention and CAC payback.

Cases and templates for a business owner

Case 1 (e‑commerce, pharma/parapharma): mROAS instead of gross RoAS. After accounting for COGS, returns and logistics, some of the “star” campaigns stopped clearing the breakeven threshold. The solution we developed at BUSINESS SITE: feed rebuild, filter by SKUs with margin above X%, new remarketing structure. Result – mROAS growth of 22% with the same revenue.

Case 2 (SaaS for B2B): ROI and CAC payback. At launch RoAS was below target, but cohort analysis showed high LTV and low churn. In our experience, focusing on ROI and payback allowed accelerating growth while preserving cash flow: scaling with control of NDR (net dollar retention) and upsell.

Case 3 (omnichannel retail): blend‑ROAS + ROI. Consolidated RoAS looked stable, but MMM and holdout tests confirmed organic cannibalization. BUSINESS SITE practice validated reallocating budget to the upper‑funnel and video: quarterly ROMI increased and CAC decreased.

Calculation templates (for Excel/BI):

RoAS = Revenue_attr / Ad_Spend
mROAS = (Revenue − COGS − Discounts − Logistics − Returns) / Ad_Spend
Net ROI = (Contribution_Margin − OPEX − Capex_Amort) / (Ad_Spend + Direct_Investments)
LTV:CAC, Payback = CAC / Monthly_Contribution_per_Customer

Recommendations for KPI set:

• SMB e‑commerce: channel‑level mROAS, blended CPA, AOV, CRO, returns, ROI by promotions.
• Enterprise/omnichannel: MMM‑level ROMI/ROI, iROAS for key campaigns, ROIC for large initiatives.
• SaaS: MRR, ARR, CAC, LTV, LTV:CAC, CAC payback, retention cohorts and NDR.

Benchmarks: RoAS/ROI targets vary by industry; adjust thresholds for margin and logistics. For fashion it’s reasonable to allow for above‑average return rates, for electronics — lower margin and higher sensitivity to discounts.

Executive dashboard – which KPIs to include

Operational layer (daily/weekly):

• RoAS by channel, margin‑adjusted RoAS, blended CPA, CPC/CPM, AOV, CR (CRO).
• Returns and logistics, frequency, creative fatigue, spend vs cap.

Strategic layer (monthly/quarterly):

• ROMI/ROI, LTV, CAC, LTV:CAC, payback period, EBITDA contribution.
• MMM estimates of channel contributions, iROAS by experiments, ROIC by projects.

Alerts and scenarios:

• mROAS below threshold for 3 days → auto‑lower bid by X% and alert to the need for a new creative.
• CAC payback exceeded target → pause scaling and analyze retention.
• iROAS < 1 on holdout → reallocate budget to channels with better uplift.

Formulas and automation of calculations

Data collection and ETL:

• Sources: advertising APIs, CRM/ERP, payment systems (PrivatBank/Monobank), web analytics, marketplaces (Rozetka, Prom.ua), logistics (Nova Poshta).
• ETL: regular loading into a data warehouse, mapping to a unified channels/campaigns directory, deduplication of orders.

Formulas:

RoAS = Revenue_attr / Ad_Spend
mROAS = (Revenue − COGS − Discounts − Logistics − Returns) / Ad_Spend
ROMI = (Incremental_Profit − Marketing_Cost) / Marketing_Cost
ROI = (Net_Profit − Total_Investment) / Total_Investment
iROAS = Incremental_Revenue / Incremental_Spend
LTV:CAC = LTV / CAC
Payback = CAC / Monthly_Contribution

Implementation checklist:

• Data validation and a single source of truth.
• Choice of attribution model and setup of server-side tracking/conversion API.
• Implementation of margin-adjusted RoAS and breakeven thresholds.
• Running incrementality tests on key channels.
• Setting up dashboards and alerts, aligning with finance on ROI/ROIC.

Example SQL for an mROAS mart by channel (simplified):

WITH orders AS (
  SELECT
    o.order_date::date AS dt,
    o.channel,
    SUM(o.revenue) AS revenue,
    SUM(o.cogs) AS cogs,
    SUM(o.discounts) AS discounts,
    SUM(o.logistics_cost + o.returns_cost) AS logistics_returns
  FROM fact_orders o
  WHERE o.status IN ('paid','delivered')
  GROUP BY 1,2
),
ad AS (
  SELECT
    a.spend_date::date AS dt,
    a.channel,
    SUM(a.spend) AS ad_spend
  FROM fact_ad_spend a
  GROUP BY 1,2
)
SELECT
  o.dt,
  o.channel,
  revenue,
  ad_spend,
  (revenue - cogs - discounts - logistics_returns) AS contribution,
  CASE WHEN ad_spend>0
       THEN (revenue - cogs - discounts - logistics_returns)/ad_spend
       ELSE NULL END AS mroas
FROM orders o
LEFT JOIN ad USING (dt, channel)
ORDER BY o.dt, o.channel;

Automation: periodic reports, scenario modeling of ROI, rules for bid strategies, predictive analytics for forecasting revenue and operational capacity.

Questions from business owners

Brief conclusions and practical next steps

RoAS и ROI не конкурируют, они дополняют друг друга. RoAS и особенно mROAS позволяют оперативно управлять эффективностью рекламных расходов, быстро тестировать гипотезы и масштабировать каналы. ROI, ROMI, NPV/IRR и ROIC задают стратегию: куда направить капитал, как сбалансировать верхнюю и нижнюю воронку, какой медиа‑микс даст максимальную чистую прибыль.

Recommended action plan:

1. Привести данные в порядок: единый data warehouse, ETL из рекламных API, CRM и платёжных систем.
2. Внедрить margin‑adjusted RoAS и пороги безубыточности с учётом COGS, логистики и возвратов.
3. Настроить ROI‑отчёты, включающие OPEX и стоимость капитала; добавить MMM или incrementality там, где много ассистов.
4. Запускать incrementality‑тесты перед существенным масштабированием, обновлять пороги по результатам uplift‑оценок.
5. Собрать executive dashboard с оперативным и стратегическим слоями и правилами алертов.