apships.com
Growth & performance marketing consultant · Delhi, India · US and UK clients

Abhishek Pal ships.

Most recently I was AVP Marketing at Internshala, where I cut candidate CAC 4x by fixing Performance Max signals. Before that, five years in growth at Naukri, India's largest jobs platform. Today I work directly with founders and marketing leads on three things: finding and fixing what leaks paid budget on Google and Meta, making conversion tracking tell the truth, and winning back users who have gone quiet. Fixed scope, fixed price, and I do the work myself.

Delhi, India
Calls in your morning. Delivery in mine.
4x
Candidate CAC cut, Rs.160 to Rs.36 against a Rs.120 target, by fixing Performance Max signals.
Internshala · 2025 to 2026
USD 2.5M+
A year (Rs.25 Cr) in marketing budgets managed across 20+ channels.
Naukri · 2021 to 2024
230K
Dormant users won back by an automated reactivation engine, at 23% lower cost, across 110+ cohorts.
Naukri · 2019 to 2021 · Platinum award
7hrs/wk
Of manual audit work saved by the negative-keyword auditor I built. In daily use.
Python · LLM APIs
Part 1

The results

Everything here happened. Each line names the company and the mechanism, because a number without a mechanism is only a claim.

Rs. is Indian rupees, roughly Rs.85 to 90 per US dollar.

Paid acquisition: a lower cost per customer by fixing the signal, not the budget

What happenedWhereHow
  • Candidate acquisition cost fell 4x, from Rs.160 to Rs.36, against a Rs.120 target.
    Internshala
    2025 to 2026
    I rebuilt the conversion signal Performance Max learns from: a custom Tag Manager event that fired only on a graduate's first application of the month, filtering out non-graduates and repeat applies. Then offline and enhanced conversion imports, and value-based bidding on search.
  • The user mix shifted from 30% to 70% graduates, the users the business actually wanted.
    Internshala
    Same fix. Once the algorithm optimised on the right event, it found the right people.
  • The EdTech vertical became ROI-positive for the first time. ROAS improved 1.5x.
    Internshala
    Value-based bidding on search plus Google offline conversions, so bids followed revenue rather than form fills.
  • 60% incremental sales from remarketing, validated with holdout groups.
    Internshala
    Cohort remarketing funnels measured against a control group, so the lift was real, not attributed.
  • Influencer output tripled at the same cost per lead.
    Internshala
    A 249-respondent survey of paying users showed how much influencer content had shaped their decision. That justified scaling the channel.
  • App installs doubled, at a 12% install-to-registration rate.
    Naukri
    Google App campaigns with registration, not the install, as the optimisation event.
Candidate CAC, Rs.
Before
160
Target
120
After
36

Dormant users: the cheapest growth most businesses ignore

What happenedWhereHow
  • An automated reactivation engine won back 230K dormant users at 23% lower cost, across 110+ cohorts.
    Naukri
    2019 to 2021
    I extended the platform's job-recommendation engine out to Meta and Google, so each dormant cohort saw personalised win-back ads while browsing off-platform. It won the company's Platinum award.
  • A separate, earlier manual pilot re-engaged 210K users at 4.5x acquisition-cost efficiency: Rs.17 per reactivated user versus Rs.80 to acquire a new one.
    Naukri
    Hand-built campaigns per cohort. It proved the economics before the engine was built.
  • An in-house lifecycle platform (email, SMS, WhatsApp, push) reaching 2.1M users, with Rs.45L of efficiencies in its first year.
    Naukri
    Built during COVID budget cuts, equivalent to a paid marketing-automation setup.
Cost per user, Rs. (manual pilot)
New user
80
Reactivated
17

Scale and range

What happenedWhere
  • Managed USD 2.5M+ a year (Rs.25 Cr) in marketing budgets across 20+ channels.
    Naukri
  • Launched four channels from scratch: Snapchat, LinkedIn, Truecaller and a GeeksforGeeks partnership.
    Naukri
  • Held brand search at -5% year on year while the category fell 16%.
    Naukri
  • Built a Rs.2.5 Cr revenue line from zero across four new states, hiring and training a nine-person team, with Meta and Google lead generation set up from nothing.
    Alrick Solar
    co-founder, 2024 to 2025

What this means for your account

Three kinds of result are on the table, and the mechanism behind each is one I have run before.

  • A lower cost per order or lead, without more budget. Most "our ads don't work" problems are signal problems. The platform is optimising on the wrong event, or on an event it cannot see. Fix the signal and the same budget buys better customers.
  • Revenue from users you already paid for. If you have a real user base going quiet, the dormant segment is usually the cheapest growth available. Almost nobody builds the machine to go and get it.
  • Numbers you can trust. A tracking setup reconciled to your actual orders, so every decision after it is made on real data.
What I put in writing is what I control: the waste I find and the accuracy of the tracking I fix. I never guarantee your ROAS, and I would not trust anyone who does.
Ship log

The tools I built

Because the existing ones were too slow. Python over LLM APIs.

  • Negative-keyword auditor
    Reads search-term reports and flags the negative-keyword gaps a human reviewer misses, returned as a ranked list of wasted spend.
    Python · LLM APIs · Google Ads exports
    About 7 hours of manual audit work saved a week. Rs.10 to 20k of wasted spend flagged a month. In daily use.
  • Brand-listening dashboard
    Replaced a paid social-listening tool for the team that monitors brand mentions.
    Python · Streamlit · LLM APIs
    About 60% less manual effort for the team that used it.
  • Lead-flow pipeline
    Sourcing, enrichment, scoring, and outreach staged for a human to send.
    Apify · Apollo · Clay · Python · LLM scoring
    Runs my own outbound. Nothing auto-sends, by design.
I bring the auditor to every paid-media audit. It comes with the work, not as a separate invoice.
Part 2

What I sell

Four fixed-scope engagements and one ongoing format. Each one starts with diagnosis and ends with a document you own.

  • Paid-efficiency audit

    Google and Meta
    Five working days from access
    For
    Brands spending on Google or Meta with a leak they can feel but cannot locate.
    I review
    • Google Ads: Performance Max versus Search allocation, brand-term defence, category head-term coverage, search-term and negative-keyword waste, Shopping and catalog health, conversion-signal quality and value-based bidding.
    • Meta: account structure, prospecting versus retargeting split, creative fatigue, catalog setup.
    • The on-site funnel, ad to landing page to checkout: promo mechanics and their margin cost, mobile load, checkout friction.
    • Measurement: GA4 and Tag Manager event integrity, and what the numbers can and cannot be trusted to say.
    You get
    A prioritised fix list, roughly ten pages, each item with the observed issue, the fix, and the projected saving or lift. A 30-minute live walkthrough. The findings are yours to implement with or without me.
    Access
    Analyst or read-only access to Google Ads, Meta Business Manager, GA4 and Tag Manager. Never a password.
  • Conversion-tracking fix

    Five working days for the standard stack
    Longer if a server-side build is in scope
    For
    Accounts where the platforms are learning from bad or missing data. Symptoms: conversions in Google Ads that do not match your order system, a low Meta Event Match Quality score, a purchase event that fires twice or not at all, or a UK cookie banner that blocks the tags before anyone answers it.
    I do
    Design the event model (which events, deduplication rules, match keys, values). Build the Tag Manager configuration and the Google-side conversion imports (offline and enhanced conversions) myself. Write the specification for the Meta Conversions API and any server-side piece, for your developer to build. Then reconcile the result against your order data and sign it off.
    You get
    A working, documented tracking setup, a reconciliation showing measured versus actual conversions, and a one-page runbook so the next person does not break it.
  • Reactivation and lifecycle sprint

    Audit first
    Build only if the audit earns it
    For
    Any business with a sizeable user or customer base that goes quiet: apps, marketplaces, subscriptions, e-commerce with a large one-time-buyer list.
    Step one
    The reactivation audit. Map the dormant base by cohort, find the cheapest path to win each back, and hand over a three-page teardown with the five highest-impact fixes and a 30-minute readout.
    Step two
    The build, only if the audit earns it. Win-back across your owned channels (email, SMS, WhatsApp, push) and paid channels (Meta and Google custom audiences), measured against a holdout group so incrementality is measured, not assumed. Built the same way I built the engine at Naukri.
  • Lead-flow build

    A running system, documented
    Every account in your name
    For
    Founder-led B2B companies where the founder still does outbound personally and there is no SDR.
    I build
    The pipeline end to end. Sourcing and enrichment through Apify, Apollo and Clay. LLM scoring that ranks fit before a human reads a row. Deduplication and a kill list, so a ruled-out prospect never resurfaces. Outreach staged for a human to send, because nothing should auto-send under your name.
    You get
    A running system, the documentation, and every account in your name.
  • Ongoing senior oversight

    Month to month
    No lock-in
    After
    Any of the above. The natural next step is month-to-month senior oversight: a weekly cadence with your team, the measurement setup maintained, an experiment roadmap, and a monthly report in plain language. Scoped to what the audit surfaced.
Prices are fixed per engagement and quoted after a 20-minute call, once I know your spend, your stack and where it hurts.
Part 3

How I work

The same process on every engagement, and the same checklist on every account.

The process, for any engagement

  1. A 20-minute call.
    You tell me the spend, the stack and the symptom. I tell you whether I can help and what it costs. If I cannot help, I say so on the call.
  2. Access, 30 minutes of your time.
    Read-only or analyst access to the ad accounts, analytics, tag manager and, where relevant, the store or CRM. I work inside your accounts. I never take passwords, and everything stays in your name.
  3. Five working days.
    I run the account and the site through the same checklist every time, then read every result myself. No juniors. No hand-off.
  4. The readout.
    A live walkthrough and a written memo: the three to five fixes ranked by recoverable monthly spend, in the order to do them, with what each needs from your team.
  5. After.
    You own every document, script and account. Implement it yourself, with your team, or with me month to month.

What I check in a paid-media audit

This is the core of the checklist. It is the same list on every account, which is why the audit takes five days and not five weeks. Every item in the memo carries the observed evidence, the fix, and the projected monthly saving or lift.

Signal and measurement
  • Which conversion event each campaign optimises on, and whether that event is the one the business wants
  • Duplicate, missing or mis-valued conversions: GA4, Google Ads and Meta counts reconciled against the order system
  • Enhanced conversions and offline conversion imports: present, matching, and feeding value
  • Meta Event Match Quality, and deduplication between pixel and server events
  • Tags firing before consent for UK visitors, and the site's response to the browser's Global Privacy Control signal for US visitors
  • Click IDs surviving the path from ad to form to CRM
Where the money leaks
  • Search-term waste and negative-keyword gaps, run through my auditor
  • Brand-term spend while ranking first organically
  • Performance Max cannibalising Search and brand traffic
  • Placement defaults switched on without a decision: Audience Network, Display expansion, Search partners
  • Prospecting and retargeting audience overlap
  • Shopping feed errors, disapprovals and missing attributes
Creative and structure
  • Creative fatigue by frequency and cost trend
  • Near-duplicate creatives that the platform treats as one
  • Campaign structure against the spend level: too fragmented to learn, or too consolidated to control
  • Budget allocation across channels against marginal cost, not average cost
The path after the click
  • Landing page load on mobile, on a real connection
  • Message match from ad to page
  • Promo mechanics and what they cost in margin
  • Checkout friction and abandoned-cart handling

Tools

Google Ads (Search, Performance Max, Shopping, App campaigns), Meta Ads, GA4, Google Tag Manager, Looker Studio, Tableau, SQL, DV360, AppsFlyer, MoEngage, Apify, Apollo, Clay, Python and LLM APIs. Where I do not know a tool in your stack, I say so before we start.

Time zones and availability

I am in Delhi. I take calls in your morning and deliver in mine, so you usually have a reply waiting when you start work.

Your hoursMy clock (IST)
UK office hours, 9 am to 5 pm1:30 pm to 9:30 pm
US Eastern, 9 am to 1 pm6:30 pm to 10:30 pm

Paperwork and payment

Invoices in USD or GBP. Fifty percent to start, fifty on delivery. US clients get a W-8BEN. UK business clients account for VAT under the reverse charge, so no VAT appears on the invoice. A short engagement letter covers scope, confidentiality, and IP, which transfers to you on payment. If I touch customer-level data, I sign your data-processing agreement.

Part 4

Who this is for

Good fit

US and UK e-commerce and lead-generation brands with active spend on Google or Meta, a founder or a small marketing team, and a specific problem: a cost per order that crept up, conversions that do not match orders, a Performance Max account nobody fully trusts, or a large customer list nobody emails. Also founder-led B2B companies where outbound still sits with the founder, and any business with a big user base that has gone quiet.

Not a fit

Companies that want an agency to run everything day to day with a team behind it. Anyone who needs a guarantee on ROAS. Enterprise marketing organisations with the specialists already in-house.

Part 5

About me

Most recently I was AVP Marketing at Internshala, where I cut candidate CAC 4x by fixing Performance Max signals. Before that, five years in growth at Naukri, India's largest jobs platform. In between, I co-founded the North Uttar Pradesh branch of a solar company and built its revenue line from zero. Eight years in commercial roles, six-plus in growth. IIFT MBA, B.Tech in IT. Since April 2026 I have been building AI marketing tools independently and consulting.

Two things about how I work that the numbers do not show. First, I do the work myself. What you read in the memo is what I saw in the account. Second, I distrust attribution by default. Where a result can be measured against a holdout, I measure it that way, and I say when it cannot.

  • Apr 2026 · present
    Independent · AI marketing tools and consulting
  • Jun 2025 · Apr 2026
    AVP Marketing · Internshala
    Led a 6-member team across performance, influencer and brand for three verticals. Hands-on SEM and Performance Max.
  • Jun 2024 · May 2025
    Co-founder & CMO · Alrick Solar
    North Uttar Pradesh branch from zero. Nine hires. Rs.2.5 Cr revenue line across four new states.
  • Apr 2021 · Jun 2024
    Senior Marketing Manager · Naukri.com (Info Edge)
    Budgets, brand campaigns, 20+ channels, direct DV360 seat, app marketing, board-level competitive intelligence.
  • Apr 2019 · Apr 2021
    Marketing Manager · Naukri.com (Info Edge)
    Lifecycle and reactivation. The automated reactivation engine and the in-house lifecycle platform.
  • Education
    MBA, IIFT New Delhi · B.Tech IT, MIET Meerut
Contact

Start with a call

Twenty minutes. Spend, stack, symptom.

You tell me the spend, the stack and the symptom. I tell you whether I can help and what it costs. If I cannot help, I say so on the call.

Replies within a working day. Every number on this page is one I can walk you through.