How to Forecast UA Spend (2025 Framework)

Most UA forecasts fail because they start with budget, not outcomes.

Teams allocate $50K to Facebook, $30K to Google, and $20K to TikTok without modeling what those budgets will actually deliver in users, revenue, or payback.

The result: budgets that look precise but bear no relationship to business goals.

A functional UA forecast starts with growth targets and works backward through conversion funnels, channel performance, and cost benchmarks to calculate required spend.

Here's how to build forecasts that actually inform decisions.

The Two Forecasting Approaches

There are two primary methods for forecasting UA spend, and the best approach uses both.

Top-Down Forecasting

Top-down starts with revenue or user growth targets and works backward to calculate required marketing spend.

The process:

1. Set revenue targets (e.g., $500K MRR by Q4)
2. Calculate required users based on conversion rate and ARPU
3. Determine traffic needs based on funnel conversion rates
4. Model budget required to generate that traffic at target CPI

This approach ensures your spending aligns with business objectives, but it assumes channels can deliver the volume you need at acceptable costs.

Bottom-Up Forecasting

Bottom-up starts with channel capacity, performance benchmarks, and builds up to total user volume.

The process:

1. Assess each channel's realistic capacity (budget, inventory, targeting limits)
2. Apply historical or industry CPI benchmarks
3. Calculate total users you can acquire within budget constraints
4. Model resulting revenue based on conversion and LTV assumptions

This approach grounds your forecast in channel realities, but it may not deliver the growth your business requires.

The Hybrid Approach

The most accurate forecasts combine both:

• Use top-down to set growth targets and validate feasibility
• Use bottom-up to model channel-specific performance and constraints
• Identify gaps where targets exceed capacity and build strategies to close them

Step-by-Step Forecasting Framework

Here's a practical framework for building UA forecasts in 2025.

Step 1: Define Your Growth Targets

Start with clear business objectives translated into user or revenue targets.

Example:

• Revenue target: $300K MRR by end of Q2
• Current MRR: $150K
• Required growth: $150K in new MRR over 90 days
• ARPU: $30/month
• Required new users: 5,000 paying users

If your conversion from install to paid is 5%, you need 100,000 installs over the quarter, or approximately 33,000 per month.

Step 2: Map Your Conversion Funnel

Model the journey from impression to revenue to understand required inputs at each stage.

Example funnel:

• Impression → Click: 2% CTR
• Click → Install: 30% install rate
• Install → Signup: 60% activation
• Signup → Paid: 5% conversion

To acquire 33,000 installs per month, you need:

• Installs: 33,000
• Clicks: 110,000 (at 30% install rate)
• Impressions: 5,500,000 (at 2% CTR)

Your forecast now has traffic targets, not just budget numbers.

Step 3: Allocate by Channel

Distribute your traffic and budget targets across validated acquisition channels based on historical performance and strategic priorities.

Channel allocation example (monthly):

Channel	Budget	Expected CPI	Projected Installs
Apple Search Ads	$18,000	$1.50	12,000
Facebook/Instagram	$30,000	$4.00	7,500
Google App Campaigns	$20,000	$2.60	7,700
TikTok Ads	$15,000	$2.90	5,200
Organic (ASO)	$2,000	—	500
Total	$85,000	$2.58 avg	32,900

This allocation is based on:

• Historical CPI performance by channel
• Current market benchmarks (see Article 1 for 2025 CPI data)
• Channel budget minimums and testing capacity

Step 4: Account for Seasonality

UA costs and conversion rates fluctuate based on time of year, competitive dynamics, and platform changes.

Common seasonal patterns:

• Q4 (Oct-Dec): CPIs increase 15-30% due to holiday ad competition
• January: Strong user intent but high competition from New Year's resolution apps
• Summer (Jun-Aug): Gaming and entertainment see increased activity; B2B apps slow
• Back-to-school (Aug-Sep): Education and productivity apps see seasonal lift

Adjust your monthly forecasts to reflect these patterns:

Q1 forecast with seasonality:

Month	Base Budget	Seasonal Adjustment	Final Budget
January	$85,000	+10% (high intent)	$93,500
February	$85,000	0% (baseline)	$85,000
March	$85,000	-5% (efficiency gain)	$80,750

Your forecast should also account for expected CPI changes due to seasonal competition.

Step 5: Build Multiple Scenarios

In 2025, with evolving platform policies and economic uncertainty, single-point forecasts create false precision.

Build three scenarios to prepare for variable outcomes:

Conservative scenario (70% confidence):

• CPI increases 20% vs. baseline
• Conversion rates decline 10%
• Required budget: $102,000/month for same volume

Base scenario (50% confidence):

• Performance matches historical averages
• Required budget: $85,000/month

Aggressive scenario (30% confidence):

• CPI improves 15% through optimization
• Conversion rates increase 10% from product improvements
• Required budget: $68,000/month for same volume

This scenario-based approach helps you plan for contingencies and communicate risk to stakeholders.

Step 6: Validate Against Historical Performance

Compare your forecast to actual historical data to reality-check assumptions.

Key questions:

• Is your projected CPI within 20% of recent actuals?
• Are conversion rate assumptions grounded in real funnel data?
• Do channel allocations reflect proven performance, or are you over-indexing on untested bets?
• Have you accounted for known changes (platform updates, policy shifts, competitive moves)?

If your forecast deviates significantly from historical performance, document why and test assumptions quickly once spending begins.

Step 7: Monitor and Update Monthly

The most accurate forecasts are living documents that evolve with real-world data.

Monthly review process:

1. Compare actual spend, CPI, and installs to forecast
2. Calculate variance and identify root causes
3. Update assumptions for upcoming months based on trends
4. Adjust channel allocation to optimize performance

Example variance analysis:

Metric	Forecast	Actual	Variance	Insight
Total Spend	$85,000	$88,200	+3.8%	TikTok required higher CPMs
Total Installs	32,900	31,500	-4.3%	Facebook CVR down 12%
Blended CPI	$2.58	$2.80	+8.5%	Shift budget to ASA

This iterative approach compounds accuracy over time as your assumptions align with market reality.

Common Forecasting Mistakes

Starting with budget instead of goals: If you begin with "we have $100K to spend," you're optimizing for spend, not outcomes. Start with required growth and calculate the budget needed to achieve it.

Ignoring funnel stages beyond install: Forecasting installs without modeling activation, retention, and monetization creates a false sense of progress. Forecast through to revenue.

Assuming linear scaling: Most channels show diminishing returns as you increase spend. Your first $10K on Facebook may deliver $2.50 CPI, but the next $10K might cost $3.20 CPI. Model saturation curves, not straight lines.

Overlooking experimentation budget: Allocate 10-15% of your forecast to testing new channels, audiences, and creatives. Forecasts based purely on existing strategies decay as markets shift.

Forgetting platform fees and fraud: Budget for 3-5% fraud or invalid traffic and factor in platform fees (e.g., Apple's 30% IAP cut, Google's 15-30% fees) when modeling revenue impact.

Key Metrics to Track

Your forecast should include more than spend and installs. Track these metrics monthly:

• Blended CPI: Total spend ÷ total installs across all channels
• CPI by channel: Platform-specific cost efficiency
• Install-to-activation rate: Percentage of installs that complete onboarding
• Activation-to-paid conversion: Percentage that monetize
• Payback period: Months to recover CAC from user revenue
• ROAS: Revenue generated per dollar spent (track at 30, 60, 90 days)

These metrics allow you to optimize not just cost per install, but cost per valuable user.

Tools and Templates

Forecasting tools:

• Google Sheets or Excel with scenario modeling
• Financial planning software (Mosaic, Drivetrain, Abacum)
• Attribution platforms (AppsFlyer, Adjust, Singular) with forecasting modules

Data sources:

• Historical campaign performance from ad platforms
• ASO tools (Sensor Tower, AppTweak) for organic forecasts
• Analytics platforms (Amplitude, Mixpanel) for conversion funnel data
• Industry benchmarks (see Article 1 for 2025 CPI benchmarks)

Final Considerations

UA forecasts are less about predicting the future and more about creating a framework for making decisions under uncertainty.

The goal isn't perfect accuracy. It's building a model that:

• Aligns spending with growth objectives
• Surfaces assumptions that need testing
• Enables fast adjustments as performance data arrives
• Communicates risk and opportunity to stakeholders

Start with your growth targets. Work backward through funnels and channels to calculate required spend. Build scenarios to account for variability. Update monthly based on actuals.

The best forecast is the one you can iterate.

FAQs

How accurate should UA forecasts be?

Aim for forecasts within 15-20% of actuals in stable markets. New channels or untested strategies may vary by 30-40% until you gather sufficient data. Monthly reviews and adjustments improve accuracy over time.

What's the difference between top-down and bottom-up forecasting?

Top-down starts with revenue targets and works backward to required traffic and spend. Bottom-up starts with channel capacity, CPI benchmarks, and builds up to total user volume and revenue. The most accurate forecasts combine both approaches.

How far in advance should I forecast UA spend?

Build detailed monthly forecasts for the next quarter and directional estimates for 6-12 months. Update quarterly as you gather performance data and market conditions shift. Rolling 3-month forecasts provide the best balance of accuracy and planning value.

Should I forecast organics separately from paid UA?

Yes. Organic and paid channels have different cost structures, attribution windows, and growth dynamics. Forecast them separately, then combine for total user acquisition projections.

What if my actual performance deviates significantly from forecast?

Investigate variance root causes immediately. If structural (e.g., CPI increased across all channels), update your assumptions and adjust future forecasts. If tactical (e.g., one campaign underperformed), optimize that specific element and monitor for trend changes.

---

UA forecasting works best when it's grounded in conversion data, flexible across scenarios, and updated as reality unfolds. Build a model you can iterate, not a static plan you defend.