How Bach AI Breaks Down Performance by Funnel Stage Without You Asking

Meta Ads Manager shows everything at one altitude. Spend, CPM, CTR, ROAS — all averaged across cold prospecting, warm retargeting, and bottom-funnel abandoned-cart sequences. The number is true on paper. It is also useless for diagnosis.

Bach AI splits your account into funnel stages automatically, applies the right KPIs at each stage, and surfaces stage-specific issues. Top-funnel CPM matters; middle-funnel CTR matters; bottom-funnel ROAS matters. The single blended number hides where the real problem lives.

The Invisible Problem

A 3.2x blended ROAS could mean a balanced funnel (top, middle, bottom all healthy) or a broken funnel (bottom 6x, middle 2x, top 1x). The blended number cannot distinguish between the two — and the diagnosis matters because they need opposite fixes.

For Indian D2C, the cost of mis-diagnosis is real. Brands that don’t separate funnel stages tend to over-invest in bottom-funnel (where ROAS looks easy) and under-invest in top-funnel (where ROAS looks weak). Three months later, the bottom-funnel audiences are saturated and there’s no fresh top-funnel pipeline feeding them.

How Bach AI Splits Your Account

Every ad set gets auto-categorised by funnel stage based on its targeting and objective:

• Top-funnel — cold prospecting audiences, interest-based or broad lookalikes, optimised for engagement, view-content, or add-to-cart events.

• Middle-funnel — warm audiences who have engaged but not purchased — site visitors, video viewers, page engagers.

• Bottom-funnel — abandoned carts, initiate-checkout, view-content-no-purchase audiences with high recency.

• Customer-base — existing buyers being retargeted for repeat purchases or referral campaigns.

The Right KPIs at Each Stage

Each stage gets evaluated against the KPIs that actually matter for it:

1. Top-funnel KPIs: CPM, CTR, view-through impressions, cost-per-add-to-cart. Direct purchase ROAS at this stage is usually misleading.

2. Middle-funnel KPIs: cost-per-add-to-cart, cost-per-initiate-checkout, time-to-purchase from first impression.

3. Bottom-funnel KPIs: ROAS, cost-per-purchase, conversion rate, cart-recovery rate.

4. Customer-base KPIs: repeat-purchase rate, AOV uplift, LTV-to-CAC ratio.

Read more on [retargeting funnels beyond abandoned-cart sequences](https://www.wittelsbach.ai/post/retargeting-funnels-for-d2c-beyond-abandoned-cart-sequences) for the broader retargeting framework.

The Funnel Health Dashboard

Inside Wittelsbach AI, the Funnel tab shows each stage as its own panel with its own KPIs. The four panels share a vertical column showing the user journey from cold to repeat. A single ‘funnel health’ score (0-100) summarises whether the stages are balanced.

Stage-specific issues surface here automatically. ‘Top funnel ROAS is 1.4x — that is healthy for your category. Bottom funnel ROAS is 8x — that is unhealthy because it suggests under-investment.’ The diagnosis is on the right altitude.

Why Bottom-Funnel Over-Investment Backfires

The bottom-funnel ROAS trap is the most common pattern Bach AI surfaces in Indian D2C accounts. The trap looks like this:

• Bottom-funnel ad sets (abandoned cart, etc.) show 6-12x ROAS because they retarget high-intent users.

• Founder increases bottom-funnel budget because ‘that is where ROAS is best’.

• The audience saturates within 30-45 days because high-intent pools are inherently small.

• Bottom-funnel ROAS starts dropping. Founder is confused.

• Top-funnel is starved, no new high-intent audiences are entering. Bottom-funnel cannot recover.

The correct response: protect the top-funnel as a pipeline-builder even when its own ROAS looks weak. Bach AI’s funnel view surfaces the underlying imbalance before the bottom-funnel decay happens.

Top-Funnel Diagnosis Without ROAS

Top-funnel cold prospecting cannot be evaluated on direct purchase ROAS — the audience is too cold and the attribution window is too long. The right diagnostic metrics:

• Cost-per-add-to-cart — the cleanest signal that the audience is finding the product compelling.

• View-through to add-to-cart rate — how efficiently impressions become engaged users.

• CTR on cold audiences — the creative-fit signal.

• Cost-per-7-day-engaged-user — a custom signal Bach AI computes that captures genuine cold-audience interest.

• 14-day attribution view vs 7-day — longer windows reveal the true contribution of top-funnel.

Middle-Funnel — The Underrated Stage

Middle-funnel is the most underbuilt stage in Indian D2C. Brands either go straight from top-funnel to abandoned-cart retargeting, skipping the engagement-and-consideration window where most conversion lives.

Bach AI surfaces middle-funnel under-investment when:

• Site visitors with no add-to-cart are not being retargeted.

• Video viewers at 50%+ are not being moved to consideration.

• Page engagers without conversion are sitting in dead audience pools.

Bottom-Funnel — The Honest View

Bottom-funnel ROAS is real but limited. Bach AI tracks the diminishing returns curve so the founder knows when bottom-funnel has hit its ceiling. Typical Indian D2C ceiling: bottom-funnel cannot exceed 20-30% of total Meta spend without saturation. Past that, additional spend buys lower-quality impressions.

Customer-Base — The Highest-LTV Stage

Existing-customer retargeting is the highest-margin stage but the most-ignored in Indian D2C. Bach AI flags accounts where customer-base spend is under 5% of total — almost always an under-investment. The right level is typically 8-15% of total spend, with focused campaigns on repeat-purchase and referral.

The ₹ Impact

Across Indian D2C accounts on Wittelsbach AI in Q1 2026:

• Average funnel rebalancing impact: 14-22% improvement in 60-day blended ROAS.

• Average middle-funnel build-out unlocked: ₹40K-1.2L/month of recovered margin.

• Customer-base retargeting uplift when introduced: 11-19% increase in repeat-purchase rate.

• Time to surface a funnel-imbalance issue: same-day on Bach AI vs 3-6 months of trial and error manually.

How Wittelsbach AI Operationalises Funnel View

Funnel-stage diagnosis is the difference between a blended dashboard and a real operating picture. Bach AI auto-categorises, auto-evaluates, and auto-surfaces the imbalances that single-altitude dashboards hide. Run a free Meta Ads audit at [app.wittelsbach.ai](https://app.wittelsbach.ai).

Frequently Asked Questions

How does Bach AI auto-categorise an ad set as top, middle, or bottom funnel?

The classification uses three signals: the audience definition (cold lookalike vs warm retargeting vs cart abandoner), the optimisation event (ViewContent vs AddToCart vs Purchase), and the audience size (very small high-intent vs very large cold). The classification is automatic but reviewable — operators can override it for any ad set with non-standard structures.

What if I run hybrid ad sets that target both cold and warm in the same audience?

Bach AI flags these as ‘mixed-stage’ and surfaces them as a structural issue rather than a clean stage. Hybrid ad sets typically under-perform because the algorithm cannot optimise for two different audience types simultaneously. The recommendation is usually to split the ad set into stage-specific versions.

Should bottom-funnel ROAS always be higher than top-funnel ROAS?

Usually yes — but the gap should not be extreme. A 3x top-funnel and 4x bottom-funnel suggests a healthy, well-fed funnel. A 0.8x top-funnel and 9x bottom-funnel suggests the funnel is starving. The ratio matters more than the absolute numbers. Bach AI surfaces the ratio and flags when it suggests imbalance.

How is the funnel view different in CBO vs ABO?

In CBO, individual ad sets are still categorised by their underlying audience and event, but the budget allocation is campaign-level. Bach AI surfaces both the ad-set-level classifications and the CBO-level effective allocation, which lets operators see where Meta has been distributing the budget across stages.

Does the funnel view help with Google Ads too, or is it Meta-specific?

The framework applies to any channel with a meaningful funnel structure. Bach AI’s implementation today focuses on Meta because that is where most Indian D2C spend lives. Google Ads funnel view is on the roadmap and behaves similarly — Search at the bottom, YouTube and Display at the top, Performance Max blending across stages.