Explore

Reflections

What went well -

Planning -

CPD -

Structurally breaking down NS into L2 leakages with sizing

Helped identify / fix - Repeat offender / Manual deactivation leakage etc

Efficient escalation reporting from ops & other counterparts to identify new MOs

DS model setup for blurred logos / Illustrative images / Inspired brands etc

Alternative keyword repo setup

Execution -

Delivery speed - High impact delivery speed since prioritisation followed hitting on major NS leakage arms first by primarily indexing on recall

Reducing visibility → Solving for repeat offender → Process workflow setup → Reducing latency / feedback loop

Collaboration - Healthy inter pod collaboration. Good collaboration with other pods - legal / malls and business teams in -

Helping setup a self serve reactive QC flow for malls

Great support on P0 legal escalations (HUL / CCPA / Markscan related cleanups)

Building cadence and support for Biz to take cognisance of High OC blocks

Scope changes - No major scope changes for any of the projects

Talent density - High talent density and rigour across teams - DS / Analytics / Ops / Tech etc

What could had gone better -

Execution -

Resource and time estimation -

Net delivery estimation could had been better for CG v1

While major part of the delay was unavoidable - Layoff / Context switch / Hackathon etc

Some of the components like Devops / environment related bits could be baked into the final timeline estimate / Better proactive communication of feature timelines to leadership

Plan better communication for central platform changes - Look at other macros changing in next 1-3M period to bake into final delivery timelines

Better internal communication and expectation setup

Planning -

Anticipating and moving fast on dependencies -

Cleanup of Suraksha was a massive collaborative effort which could had been started a bit early

Better alignment of PoC with category head / top down approach

Impact of these false positives could had better evangelised with category heads at the start to get better resource allocation and focus on the net cleanup

Comprehensive list of ops check metrics / Talent hiring speed -

Agent error improvement tracking was a bit delayed that led to some inefficiencies in the process

Was rectified by changing the set of agents that were doing the dipstick, both quantity and quality wise (Increased count from 15 mediocre to 80 high quality agents)

Predictability and measurement of NS

Initial NS definition was not comprehensive enough and was updated thereafter

Other next steps -

Hygiene / Alerts for other services - Seller SSR etc

Away model setup for SF side changes as part of cataloging v2

FE / BE for other KTLOs / adjacent service maintenance -

Not evident visibility of dependencies of other services and required effort for the devs

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

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

CtrlP

) instead.