Explore

Pod review

Browsing experience - We took a goal to drive 2% NMV/Vi in this pod, we have line of sight to achieve 0.6% NMV/Vi despite headwinds in scaling key initiatives (HVFs and Parallel feeds). This is primarily driven by - (i) Hero product model rebuild [+0.6%] , along with the launch of reco re-discovery on PLPs and parallel feed usability optimisation

Additional features like interstitial filters, DS filter service rationalisation and post filtered feed ranking are in pipeline but might not be scaled before the end of cycle given timelines and will spill over in the next cycle, postponing impact realisation from these initiatives.

Parallel Feed v1 experiment resulted in a +3.5% significant uplift in Gold O/Vi and an 18% improvement in view share, but platform metrics dropped by ~0.5% due to a ~5% drop in reco views caused by scrolling issues on Android devices. Fix and experiment relaunch is planned post code freeze

Hero Product rebuild is expected to boost platform conversion by +0.6% after scaling. Fashion super portfolios saw a ~1% net conversion improvement from updated hero products, closing previous performance gaps in terms of clicks and conversion. Scale-up for additional portfolios planned pre code freeze

More Like This [PLP] launched on August 29th, showed a slight 0.1% O/V increase but a 0.18% overall drop due to a decline in V/Vi. High intent REs like Search and Wishlist saw positive O/Vi gains of 0.26% and 2.3% respectively, with a strong correlation between real estate’s CTR and conversion. Team is analysing user feedback and investigating DoD CTR drop and RE-level check metrics, with plans to relaunch the experiment post code freeze with FTUX.

In this R2R, Browsing experience set the goal to improve platform NMV/Vi by 2% by Oct’24, August goal was a 0.7% increase in incremental NMV which fell short due to -

Key intent channelisation bets like visual filters / HVF revamp didn't meet platform conversion expectation. Despite significant impact on input / increased filter usage, ODNR-based ranking lowered filtered feed’s conversion

Tech solutioning and implementation delays in DS filter service rationalisation impacted accurate and comprehensive filter mapping across feeds

Operational delays in image sourcing/vetting which slowed interstitial filter experimentation

Delayed scale-up of parallel feeds due to scroll and navigation issues on custom Android devices

Parallel feed's platform NMV impact was postponed due to the base build addressing High ASP

We are on track to achieve ~90% of August goals, with the phased hero product model revamp expected to boost platform conversion by 0.6%. Additional features like Reco re-discovery on PLPs, interstitial filters, and DS filter service rationalisation are in the pipeline to address the NMV shortfall by the end of the cycle.

Key wins -

Lightspeed revamp of hero product model, which was tested across 3 leading super portfolios, with an expected platform conversion impact of ~0.6%

Lightspeed execution and experiment launch for Reco re-discovery on PLP and browsing feeds aim to improve platform conversion

Team was able to understand key issues leading to platform downside for parallel feed and push app updates to optimise feature usability across devices

Consistent CPD within team to better understand system architecture and address issues causing a buggy browsing experience, particularly with wishlist, sort, and filters.

NMV bridge view / recovery plan -

⁠

⁠

⁠

Sub KR summary -

KR Type

KR level only

Achievement status

Target Jul'24

Jul'24 Actuals

Target Aug'24

Aug'24 Forecast

Target Sep'24

Oct'24 Forecast

Top call outs

Recovery Plan (in case of shortfall)

KR success metric

Long term trendlines (visual charts for KR level only)

Pod Name

PFS/ WS/ Deep Dive Link (optional)

KR 4: Browsing Experience (Intent Channelisation & Intent Generation) to generate 2% NMV growth by Oct’24

0.4%

0.0%

0.69%

0.2%

1.69%

1.0%

Browsing experience

Sub KR4.1: Effective intent channelisation through revamped sort / visual filters to realise 1.3% NMV growth by Oct’24

4.1.1 - Revamping HVFs (+0.3% NMV/Vi)

4.1.2 - Revamping IFs (+0.3%)

4.1.3 - Structural architectural fixes (+0.6%)

Post filter ranking

DS filter service rationalisation

Static filter screen and sort bar UX revisit

Hero product model restoration

0.3%

0.0%

0.59%

0.2%

1.24%

0.7%

Highlights -

Hero product model restored and scaled to 3 super portfolios (Men Fashion, Kurtis, Home Decor) yielding a platform conversion (O/V) uplift of ~0.2%. Expected uplift with full scale-up across SP - ~0.6%

Legacy HVFs restored across core REs, improving filter usage by 1pp with a neutral impact on platform conversion

Ongoing CPD within the team to better understand system architecture and resolve issues causing buggy browsing (wishlist, sort, and filters) -

Fixes for price sort handling with offer changes (ETA: Oct 2nd week)

Fixes to address wishlist usability issues (e.g., disappearing products)

Lowlights -

Key bets on visual filters/HVF revamp fell short of conversion expectations; despite higher filter usage, ODNR-based ranking lowered filtered feed conversions

Delays in launching the IF experiment due to operational gaps in image sourcing/vetting

Tech solutioning and implementation delays in DS filter service rationalization impacted expected pre-code freeze launch

Post filter ranking - Post-filter ranking: Extending baseline rankers on filtered feeds to improve relevance and engagement

DS filter service rationalisation - Ensuring filters displayed for a given feed are comprehensive and relevant

Incremental NMV/Vi

Browsing experience

Sub KR4.2: Intent generation - Launching new REs / touch-points to realise 0.7% NMV growth by Oct’24

4.2.1 - Integrating product feeds in significant traffic real estates without browsing to generate intent

4.2.2 - Enabling swifter High ASP browsing through parallel feeds on PDP

4.2.3 - Enabling swifter adjacent browsing for exploratory / med intent users through parallel feeds

0.1%

0.0%

0.1%

0.45%

0.3%

Highlights -

More like this launch - Lightspeed execution and experiment launch for Reco re-discovery on PLP and browsing feeds aim to improve platform conversion

Lowlights -

Delayed scale-up of parallel feeds due to scroll and navigation issues on custom Android devices - App fixed pushed to reduce reco V/Vi delta from 13% → 5%

Parallel feed's platform NMV impact is postponed due to the base build addressing High ASP

Extending parallel feed to other REs for adjacent intent levers

Improving CTR for More like this via FTUX enhancements

There are no rows in this table

⁠

Continuous problem discovery -

Problem / Opportunity discovered

Owner

Explanation/ Hypothesis

10x?

Next steps

Feedback by Strat Team

Adjacent category browsing

Adjacent category browsing is quite rampant on the platform which is evident from -

~75% users heavily browse SSCAT A and end up purchasing SSCAT B within the same session

~60% users primarily browse a given super portfolio and end up buying from some other super portfolio within the same session

~17% users order from multiple SSCATs within the same day contributing to ~35% plat OC

Yes

To be solutioned as part of scaling parallel feeds across REs

Incorrect sort and filter results are being displayed for price and discount-related S&F application

Price and discount sorting issues stem from a missed integration of rating and offer expiry event with the caching service that powers sort and filter functionality

Tech team to pick the solve, Timelines awaited

ODNR based ranking post filter application

Efforts to boost conversion by expanding visual filter coverage and optimizing filter relevance did not achieve the expected results, as conversion rates for filter-applied feeds remained relatively low.

Yes

To be prioritised by ranking tech / product

Low usability and relevance for dynamic filters

Low usability persists despite the static placement of dynamic filters on the ever visible filter bar, primarily due to relevance gaps

To be picked as part of basic filter UX revamp

There are no rows in this table

⁠

Want to print your doc?
This is not the way.

Try clicking the ··· in the right corner or using a keyboard shortcut (

CtrlP

) instead.