CRO Roadmap: Building a 90-Day Plan

A 90-day CRO roadmap converts CRO from a vague goal ("improve conversions") into a structured program with specific tests, owners, timelines, and success metrics. Without a roadmap, CRO efforts become reactive — testing what's easy or what someone in a meeting suggested, rather than what will move revenue. The most successful Indian D2C brands run CRO as a systematic program: prioritized, measured, and continuously improved. This 90-day plan gives you the structure to do the same, whether you're starting from zero or resetting an underperforming program.

Brands like Bellavita achieved an 11% CVR improvement not by running one big test — but by running a systematic test program across multiple pages over time. This is the type of program this roadmap builds.

Before You Start: Prerequisites for a Successful CRO Roadmap

1. Analytics tracking must be accurate You cannot run a CRO program without reliable baseline data. Before starting the 90-day plan:

Verify GA4 or your analytics tool is tracking purchases, add-to-carts, and key micro-conversions correctly
Confirm that Shopify's built-in analytics aligns with GA4 data (within ~5%)
Verify that COD orders are tracked correctly (COD order tracking is a common gap in Indian ecommerce analytics)

2. A/B testing tool must be installed and tested Install CustomFit.ai (or your chosen tool) and run a simple "null test" — a test with no actual variation between Control and Variant. This confirms traffic is splitting correctly and that conversions are tracking.

3. Define your primary conversion goal For most D2C Shopify brands, this is purchases (order completion rate). Set a clear primary metric before starting. If your team disagrees on what "conversion" means, resolve this before building the roadmap.

Phase 1: Audit and Baseline (Days 1–14)

Week 1: Data Audit and Funnel Mapping

Your first task is to understand where visitors drop off in your funnel and what your current baseline metrics are.

Funnel mapping template:

Stage	URL/Page	Monthly Visitors	Conversion to Next Stage	Drop-off Rate
Homepage	/	15,000	35% to Category	65%
Category Pages	/collections/*	5,250	40% to Product	60%
Product Pages	/products/*	2,100	18% Add to Cart	82%
Cart	/cart	378	65% to Checkout	35%
Checkout	/checkout	246	72% Order Completion	28%
Orders		177

Calculate your stage-by-stage drop-off. In the example above, the biggest absolute drop is at the product page stage (82% drop-off without adding to cart). This is where CRO can have the most impact.

Week 1 deliverable: Funnel mapping with stage-by-stage conversion rates and a clear identification of the biggest drop-off point.

Week 2: Qualitative Research

Data tells you where visitors drop off. Qualitative research tells you why.

Tools to use:

Session recordings (Microsoft Clarity — free, or Hotjar)
Heatmaps (same tools)
Exit surveys ("Why are you leaving today?") — use a tool like Hotjar Surveys or a simple exit popup with a single question
Customer interviews — 5 recent customers who completed a purchase + 5 who abandoned cart

What to look for in session recordings:

Are visitors scrolling past key content (reviews, descriptions) without reading?
Is there rage-clicking on elements that look clickable but aren't?
Are visitors getting to checkout and dropping off at a specific form field?
On mobile, is any element causing layout breaks or overlap?

Week 2 deliverable: Qualitative insights document — 10–15 observed friction points with session recording timestamps as evidence.

Phase 2: Build Your Test Backlog (Days 15–30)

Generating Test Hypotheses

For each friction point identified in the audit, write a hypothesis:

"If we [change X] on [page Y], then [metric Z] will improve because [reason based on qualitative data or established CRO principles]."

Example:

"If we move the review summary (star rating + top 3 reviews) above the 'Add to Cart' button on product pages, then add-to-cart rate will increase because session recordings show 40% of visitors leave without scrolling to the review section below the fold."

Generate 15–20 hypotheses before prioritizing.

Prioritizing with the PIE Framework

Score each hypothesis on three dimensions (1–10 scale):

P — Potential: How much improvement is possible? A page with 2% CVR has more potential than one already at 8%.

I — Importance: How much traffic and revenue does this page have? A product page with 5,000 monthly visitors is more important than one with 200.

E — Ease: How easy is it to implement the test? Changing headline copy is easy (score 9). Redesigning checkout flow is hard (score 3).

PIE score = (P + I + E) / 3

Rank your 20 hypotheses by PIE score. The top 5–8 are your 90-day test backlog.

Example PIE Scores:

Hypothesis	P	I	E	PIE Score
Move reviews above fold on PDP	8	9	8	8.3
Add COD badge to checkout CTA	7	9	9	8.3
Test headline on top product page	7	8	9	8.0
Free shipping progress bar in cart	8	7	7	7.3
Redesign category page grid	6	8	4	6.0

Phase 2 deliverable: Prioritized test backlog with PIE scores, written hypotheses, and estimated test duration for each.

Phase 3: Run First Experiments (Days 31–60)

Launching Tests 1 and 2

Start with your top 2 PIE-scored tests. For a team of 1–2 people, running 2 tests simultaneously is manageable. More than 3 concurrent tests typically degrades quality of analysis.

Test setup checklist:

Hypothesis written and documented
Primary metric defined
Secondary and guardrail metrics defined
Sample size calculated (use CustomFit.ai's built-in calculator or Evans Miller tool)
Test duration set
Variant built in CustomFit.ai's visual editor
SRM check scheduled for 48 hours post-launch

SRM check at 48 hours: Log into CustomFit.ai and verify traffic split is within expected range. If SRM is detected, pause the test and investigate before accumulating invalid data.

Mid-test communication: Resist the temptation to peek at results and make decisions. Send a weekly update to stakeholders showing "Test 1 is in flight — we'll have results by [date]." This manages expectations and prevents pressure to stop tests early.

Days 31–60 deliverable: 2 tests launched and running with proper documentation. SRM verified. Stakeholders updated.

Phase 4: Analyze, Roll Out, and Launch Test 2.0 (Days 61–90)

Week 9: Analyze Test Results

When your tests complete their pre-determined duration:

Decision framework:

Statistically significant positive result (95%+ confidence): Roll out the winner. Document what worked and why.
Statistically significant negative result: Document what you learned. The losing variant revealed something important about your users. Update your hypothesis model.
No significant result: Don't call a winner. Run a follow-up test with a different approach if the hypothesis seems sound, or move on to the next test in the backlog.

Result documentation template:

Test: [Name]. Hypothesis: [Hypothesis]. Result: Control CVR [X%] vs. Variant CVR [Y%]. Confidence: [Z%]. Revenue impact: ₹[N]/month. Decision: [Rollout/Discard/Iterate]. Learning: [Key insight].

Week 10: Roll Out Winners and Celebrate

Roll out winning variants to 100% of traffic in CustomFit.ai. Brief your team on the result. A winning test is worth communicating — it builds momentum and organizational support for the CRO program.

Week 10–13: Launch Second Wave

With 2 tests complete (win or lose), you have learnings. Apply them to refine the next 2–3 hypotheses from your backlog. Some hypotheses become better based on what you learned:

"Test 1 showed reviews above fold increased ATC 18% → Now test which review type works best: text review vs. photo review above fold."
"Test 2 showed COD badge increased checkout clicks 12% → Now test whether adding 'Free Returns' next to COD badge further increases completions."

This iterative refinement is how CRO programs mature. Your second wave of tests is informed by real data from your first wave.

Days 61–90 deliverable: Tests 1 and 2 complete with documented results. Winners rolled out. Tests 3 and 4 launched with refined hypotheses.

The 90-Day CRO Roadmap Template

Week	Activity	Owner	Deliverable
1–2	Analytics audit and funnel mapping	Growth Manager	Funnel drop-off report
2	Qualitative research (recordings, surveys)	Growth Manager	Friction point list
3	Hypothesis generation	Team	15–20 hypotheses
4	PIE scoring and backlog prioritization	Growth Manager	Prioritized test backlog
5	Set up testing tool, run null test	Growth Manager	Tool live and validated
6–7	Build and launch Tests 1 & 2	Growth Manager	Tests in flight
7–9	Tests running (no peeking)	Growth Manager	Weekly status updates
10	Analyze Test 1 & 2 results	Growth Manager	Test results documented
10	Roll out winners	Growth Manager	Winners live
11–12	Build and launch Tests 3 & 4 (refined)	Growth Manager	Second wave in flight
13	90-day retrospective and planning	Team	Q2 roadmap drafted

What to Track for Your 90-Day Review

At the end of 90 days, report on:

Tests run: How many started, how many completed, how many reached significance
Win rate: What percentage of tests produced statistically significant positive results (typical: 30–50%)
Cumulative CVR improvement: Total conversion rate improvement from all winning tests rolled out
Revenue impact: Monthly incremental revenue attributed to CRO winners
CRO investment: Tool cost + team hours × hourly rate
CRO ROI: (Revenue impact - Investment) / Investment × 100
Test backlog health: How many validated hypotheses are waiting for the next 90-day cycle

Tips / Best Practices

Fix tracking first — you cannot run a valid CRO program on bad data. Spend 3–5 days auditing analytics before generating any hypotheses.
Use the PIE framework rigorously — gut-feeling prioritization leads to testing what's interesting, not what's important. Score every hypothesis numerically.
Run a null test before launching real tests — verify your tool is correctly splitting traffic before you run an experiment with real stakes.
Protect test duration from stakeholder pressure — define end dates upfront and communicate them. "This test runs until March 21st" is much easier to defend than "we'll look at the data next week."
Build a test archive from day one — every test should be documented in a shared document or project management tool. This archive becomes your most valuable CRO asset over time.
Plan for festive season impact — if your 90-day roadmap overlaps with a major festive season (Diwali, Navratri, year-end), either plan festive-specific tests or account for the traffic behavior difference in your analysis.
Share results broadly — a 90-day CRO report shared with the full team builds culture around data-driven decision making and makes future CRO program expansion easier to justify.

Key Takeaways

A 90-day CRO roadmap moves from audit and baseline (Days 1–14) through backlog building (Days 15–30) to first experiments (Days 31–60) and first results with refined second wave (Days 61–90)
The PIE framework (Potential × Importance × Ease) is the most practical prioritization tool for small D2C teams with limited testing capacity
Fix analytics and verify tracking accuracy before generating a single hypothesis — bad data leads to bad decisions
Most D2C teams can run 3–6 meaningful tests in 90 days — quality over quantity
Document every test outcome including losses — your test archive is your most valuable long-term CRO asset
CRO programs typically reach meaningful ROI (3–5×) by month 3–6 when structured and run systematically