Danny Deepak
← All casesJewellery D2CBengaluruClosed

MR Jewels

Google said the store sold ₹12.56 crore in August. The whole company sold ₹1.07 crore. The real money was leaking in the cart and in returns.

−66%

showroom cost per click after one conversion-goal fix (₹38.60 → ₹13.13) Measured

Model wearing an MR Jewels bird ear cuff
Photo: mrjewels.in (client)
₹12.4 Crnet revenue in the 24 months before I took over
2.68×blended return on ad spend, July 2026: lowest in 24 months
₹1.07 CrAugust gross sales: +21.8% on July, +5.7% on the 12-month average
19.8%of gross sales reversed: ₹2.41 Cr a year
Level 1

The 60-second story

For the CEO. Read left to right.
  1. 1Problem

    Google spend rose 143% in August; gross sales rose 21.8%. Profit could not be judged: margin and Meta spend were never supplied.

  2. 2The leak

    Checkout (cart to order 13.5%), returns (19.8% of gross reversed) and tracking (Google claimed 11.8 times the company's revenue).

  3. 3How I found it

    Reconciled Google exports to Shopify, billing and the client's 24-month workbook; per-campaign daily exports pinned the defect.

  4. 4What I did

    Built brand Search, showroom PMax and cart remarketing; fixed the showroom goal; specced tracking repair, call tracking and three scale stages.

  5. 5What changed

    Showroom CPC ₹38.60 to ₹13.13; 1,356 showroom actions in August. Google spend +143.3%, orders +16.1%.

  6. 6Proof

    Google Ads campaign, daily and search-term exports; Shopify 12-month export; billing summary; client MER workbook.

  7. 7Lesson

    Reconcile Google to Shopify weekly before scaling. Budget follows measured demand. Fix the cart before buying more traffic.

Level 2

The business first

Before any channel: what they sell, to whom, and how money comes in.
Business model
D2C Shopify store (mrjewels.in) plus one Indiranagar showroom; most orders complete in third-party headless checkouts
Industry
Premium imitation jewellery (kadas, earrings, men's brooches, festive sets)
Target customer
Phone shoppers across Indian metros; four in five buyers are new customers
Persona
A phone shopper in a metro, browsing midday or 9-11PM, often a man buying festive jewellery for a woman (known audience only ~4% of impressions, directional).
Geography
Bengaluru showroom; demand spread over eight metros (45.9% of Indian sessions); Bengaluru 9.0%; 7.4% of sessions outside India
Commercial objective
Grow orders (plan documents target 50 Google orders a day by November and 100 total); profit goal unknown because margin was never supplied
How revenue is made
Online orders at AOV ~₹5,180 (prepaid and cash-on-delivery); showroom purchases close in person (value not tracked)
How they convert
Product page to cart to checkout; 75.8% of sales run through the fastrrv3 headless checkout; showroom visit as second route
01

What the customer sees: before and after

The same touchpoint on a phone, before my work and after it. Screens are schematic; every change and number comes from the account, the change log or the client report.

Looking for a jewellery store in IndiranagarLiveMeasured
Before
jewellery shop indiranagar
Sponsoredmrjewels.inMR Jewels · Shop online

Showroom visits not counted as a goal. ₹38.60 a click.

After
jewellery shop indiranagar
SponsoredMaps · MR Jewels IndiranagarVisit the showroom · Directions · Call

Showroom Performance Max with local actions as the goal. ₹13.13 a click.

Maps1,356 showroom actions in August

Calls, directions and store visits counted.

  • Showroom goal fixed 21–25 Aug.
  • Three asset groups: Store & Showroom, Bridal & Trousseau, Festive & Gifting.
02

From digital to revenue: the strongest verified chain

Verified for 6 of 7 links. The chain breaks at “Profit”, so revenue attribution is unavailable beyond that point.

Google Ads spend (billed, Aug 2026)₹10,87,899 (media ex-GST ₹9,25,231)Measuredmrjewels-august-2026-report index.html s01
Google impressions (Aug 2026)1,15,59,854Measuredmrjewels-august-2026-report index.html s01
Store sessions, all channels4,20,817 (+53.3% vs July)Measuredmrjewels-august-2026-report index.html s02
Orders, all channels1,960 (+16.1%)Measuredmrjewels-august-2026-report index.html s02
Orders traced to a channel (12 months)3.4% (708 of 21,015); Google-to-order link not provable in ShopifyMeasuredmrjewels-august-2026-report index.html s12
Net revenue (Aug 2026)₹80,21,935 (gross ₹1,06,84,206 less discounts and derived reversals)Calculatedmrjewels-august-2026-report index.html s01
ProfitChain breaks hereData not availableCOGS/gross margin never supplied; Meta spend for Aug 2026 never supplied
03

Where the money leaks: the journey

This business's own path from first search to revenue. Tap a stage for the five-step leak card: what was wrong, how I found it, the evidence, what changed, and after.

01Ad reach and budget mixImproved02Showroom visit (Indiranagar)Fixed03Brand searchFlagged04Cart to orderFlagged05Delivery and returnsNot marketing06Repeat purchaseFlagged07MeasurementFlagged08Billing and deliveryFlagged
01Ad reach and budget mixImprovedFour rupees in five sat on Shopping; Search, the cheapest layer, got one in a hundred (also 66% of the feed never shown)
Stage metricShopping 80.9% of spend at ₹1,146 per conversion; Search 1.0% at ₹73 Measuredmrjewels-google-ads-case-study docs/index.html s02; data/campaign-type-performance.csv (all-time, previous team)
  1. 1 · What was wrong

    Four rupees in five sat on Shopping; Search, the cheapest layer, got one in a hundred (also 66% of the feed never shown)

  2. 2 · How I identified it

    All-time campaign-type, search-term and landing-page exports reconciled on 17 Aug

  3. 3 · Evidence

    Shopping CPA ₹1,146 vs Search ₹73 (15.6x); 8.6 L search terms, 2,692 converted Measuredmrjewels-google-ads-case-study docs/index.html s02, s06

  4. 4 · What changed

    Built brand Search, three new PMax asset groups and a surgical negative list from 13.7 L search terms

  5. 5 · After

    New asset groups ₹20-29 per conversion vs ₹287-367 on inherited groups (new groups count showroom actions; compare with care) Measuredmrjewels-seven-weeks index.html s04

02Showroom visit (Indiranagar)FixedNo local action was set Primary, so Smart Bidding ran blind; store campaign read zero for 17 days at ₹38.60 CPC
Stage metric1,356 local actions in August on ₹27,025 (3% of budget) Measuredmrjewels-seven-weeks index.html s04
  1. 1 · What was wrong

    No local action was set Primary, so Smart Bidding ran blind; store campaign read zero for 17 days at ₹38.60 CPC

  2. 2 · How I identified it

    Asset-group report 1-17 Aug against conversion goal configuration

  3. 3 · Evidence

    CPC ₹38.60, CTR 0.69%, 122 clicks (1-17 Aug) Measuredmrjewels-seven-weeks index.html s04 'The fix that paid'

  4. 4 · What changed

    Conversion goal fixed 21-25 Aug

  5. 5 · After

    CPC ₹13.13 (-66%), CTR 2.40%, 1,241 clicks (1-25 Aug); part of the 543 to 1,356 jump is a counting change Measuredmrjewels-seven-weeks index.html s04; hist caveats

03Brand searchFlaggedBrand budget raised ₹2,000 to ₹3,100 a day against a spec cap of ₹150-200; surplus went to broad match on non-brand clicks
Stage metricBrand: 207 terms, 1.3% of spend, 24.8% of conversions at ₹46 (all-time) Measuredmrjewels-google-ads-case-study docs/index.html data/search-term-intent-buckets.csv
  1. 1 · What was wrong

    Brand budget raised ₹2,000 to ₹3,100 a day against a spec cap of ₹150-200; surplus went to broad match on non-brand clicks

  2. 2 · How I identified it

    24 Aug bid-strategy pre-flight; brand demand measured at ~₹64 a day

  3. 3 · Evidence

    Search | Brand | Bangalore ₹57,249 at ₹2,408 per conversion vs Brand_3june ₹3,093 at ₹36 Measuredmrjewels-august-2026-report index.html s06

  4. 4 · What changed

    Spec to pause Brand Broad 01, cap brand-exact CPC ₹8-12, rebuild on phrase (handed over)

  5. 5 · After

    Data not availableNot confirmed built before close (SW s15 shows 'FIX' status)

04Cart to orderFlagged85 of 100 carts never buy; 1,43,983 sessions added to cart in a year, 1,23,133 left
Stage metricSession-to-cart 5.45%; cart-to-order 13.5% (earlier 16.8%; periods unstated) Measuredmrjewels-seven-weeks index.html s06 findings
  1. 1 · What was wrong

    85 of 100 carts never buy; 1,43,983 sessions added to cart in a year, 1,23,133 left

  2. 2 · How I identified it

    24-month blended review, Shopify funnel counts

  3. 3 · Evidence

    At 25% cart-to-order the store would do 98.6 orders a day on today's traffic Calculatedmrjewels-seven-weeks index.html s14 'cart to order' chart

  4. 4 · What changed

    Launched Demand Gen cart remarketing (0-7 day audience, 3 ads)

  5. 5 · After

    ₹11,971 spent, 3,118 clicks, 0 purchases by 2 Sep; 92 view-through conversions Measuredmrjewels-seven-weeks index.html s13 item 7

05Delivery and returnsNot marketingLargest cost in the business was unmonitored; one checkout reverses at 3.6x the native store; pendants reverse 40-63%
Stage metric₹2.41 Cr reversed in 12 months, 19.8% of gross sales Calculatedmrjewels-august-2026-report index.html s05 (monthly values derived from 12-month rate)
  1. 1 · What was wrong

    Largest cost in the business was unmonitored; one checkout reverses at 3.6x the native store; pendants reverse 40-63%

  2. 2 · How I identified it

    Shopify sales by channel and by product, 12 months

  3. 3 · Evidence

    fastrrv3 23.8% on 75.8% of sales vs 6.7% native Online Store; ₹1,19,134 of August ad spend on SKUs with 30%+ reversal history Measuredmrjewels-august-2026-report index.html s05

  4. 4 · What changed

    Flagged for client ops with prepaid-incentive test on 40%+ pendant lines; low-reversal kadas proposed as own campaign

  5. 5 · After

    Data not availableOps outcome after handover not recorded

06Repeat purchaseFlaggedAcquisition treadmill: repeat decays to ~1% by month 6
Stage metric79.6% new customers (14,534 new, 3,733 returning); month-1 repeat 12.6% Measuredmrjewels-august-2026-report index.html s12
  1. 1 · What was wrong

    Acquisition treadmill: repeat decays to ~1% by month 6

  2. 2 · How I identified it

    Shopify customers report, 12 months (first new/returning split the business had)

  3. 3 · Evidence

    Returning customer rate 23.51% (Shopify headline) vs 20.4% returning of 18,267 buyers Measuredmrjewels-august-2026-report index.html s12

  4. 4 · What changed

    Email and WhatsApp lifecycle flows for first-time buyers (in build)

  5. 5 · After

    Data not availableNo result recorded

07MeasurementFlaggedStand_14Jan booked ₹12.33 Cr on 1-4 Aug on ₹53,896 spend; purchase counted twice; add-to-cart value bug; Google conversions also drifted from store orders in Jul 2026
Stage metricGoogle reported ₹12.56 Cr value in August; company gross sales ₹1.07 Cr Measuredmrjewels-seven-weeks index.html s07; mrjewels-august-2026-report index.html s03
  1. 1 · What was wrong

    Stand_14Jan booked ₹12.33 Cr on 1-4 Aug on ₹53,896 spend; purchase counted twice; add-to-cart value bug; Google conversions also drifted from store orders in Jul 2026

  2. 2 · How I identified it

    Per-campaign daily exports; Shopify revenue and billing used as control; 2 Sep conversion-action review

  3. 3 · Evidence

    Jul 2026: Google conversions +37.0% on cost -19.3% while Shopify orders -5.3% and gross sales -14.8% (my calc, 12-month series) Calculatedclient monthly history google_conversions, google_cost_inr, shopify_orders, shopify_gross_sales

  4. 4 · What changed

    Tracking-repair spec (3 defects, 1 false alarm retired) and weekly Google-vs-Shopify reconciliation handed over

  5. 5 · After

    Data not availableRepair not confirmed built; no post-fix reconciliation in sources

08Billing and deliveryFlaggedManual ₹15,000 netbanking top-ups on a ₹10.88 L/month account; account went dark twice in August
Stage metricAvailable funds ₹18,571 on 1 Sep vs ₹33,000/day run rate Measuredmrjewels-august-2026-report index.html s08
  1. 1 · What was wrong

    Manual ₹15,000 netbanking top-ups on a ₹10.88 L/month account; account went dark twice in August

  2. 2 · How I identified it

    Daily spend series: 10 Aug spend ₹533 vs ₹14,900/day before

  3. 3 · Evidence

    About ₹32,000 of delivery lost on 10 Aug Estimatedmrjewels-august-2026-report index.html s01

  4. 4 · What changed

    Auto-pay, backup card and three-day balance alert recommended

  5. 5 · After

    Data not availableNot recorded as done

04

Before, what I found, action, after

Verified numbers only. Where an after number does not exist yet, it says so.

Showroom CPC (PMax Store)
Before₹38.60 (1-17 Aug) Measuredmrjewels-seven-weeks index.html s04
What I found

No local action set to Primary; bidding bought anonymous clicks for 17 days

Action

Fixed conversion goal 21-25 Aug

After₹13.13 (1-25 Aug); ₹10.13 (18-25 Aug) Measuredmrjewels-seven-weeks index.html s04
Showroom local actions
Before543 in 25 months Measuredmrjewels-seven-weeks index.html s04; playbooks/03 s22
What I found

Showroom got 0.22% of lifetime budget; directions ₹32.57 each

Action

Store campaign plus goal fix; share of budget raised

After1,356 in Aug 2026 (part is a counting change) Measuredmrjewels-seven-weeks index.html s04
Showroom share of budget
Before0.22% (lifetime) Measuredmrjewels-seven-weeks index.html s04
What I found

Under-funded cheapest high-intent signal

Action

PMax Store launched 1 Aug

After2.9% (Aug 2026, ₹27,025) Measuredmrjewels-seven-weeks index.html s04
Cost per conversion, new vs inherited PMax asset groups
Before₹287-367 (inherited Catch All and Open) Measuredmrjewels-august-2026-report index.html s14
What I found

Inherited groups at Average ad strength

Action

Wrote Store, Bridal and Festive asset groups

After₹20-29 (new groups; they count showroom actions) Measuredmrjewels-august-2026-report index.html s14
Orders per month vs Google billed
Before1,688 orders on ₹4,47,196 (Jul 2026) Measuredmrjewels-august-2026-report index.html s02
What I found

Value signal unverified; scale decided on defective data

Action

Scaled Google spend in August (S_Tofu, CatchAll, Stand_14Jan, brand)

After1,960 orders on ₹10,87,899 (+16.1% orders, +143.3% spend); incremental 2.99x gross, 2.24x net Measuredmrjewels-august-2026-report index.html s02
05

What the history shows

Patterns found by re-reading the full history, not just the months I ran. Each is labelled: measured pattern, correlation, anomaly or seasonality.

Correlation, not causation

February 2026 already showed the August 'wider, not better' pattern; orders sit near a 2,000 a month ceiling

Jan to Feb 2026: sessions +39.5%, orders -4.4%, conversion rate 0.687% to 0.471%. Aug 2026 landed on the same point: 4,20,817 sessions, 1,960 orders, 0.466%, versus Feb's 4,18,236, 1,970, 0.471%. Orders stayed in 1,642-2,060 from Oct 2025 while sessions swung 2,36,079-4,20,817. AOV and conversion move opposite ways: AOV +14.0% Sep 2025 to Aug 2026, conversion -42.0%, gross sales per session ₹38.6 to ₹25.4. Aug 2026 vs Aug 2025: orders +58.6% (1,236 to 1,960) on Google billed x6.48 (₹1,67,886 to ₹10,87,899), about ₹1,271 of extra Google billing per extra order, Google only; Meta Aug 2026 unknown.

The numbers

Sessions vs conversion rate r = -0.93 (n=12); sessions vs orders r = +0.74; log-log slope 0.37 (1% more sessions, ~0.37% more orders); AOV vs conversion r = -0.55; Feb 2026 and Aug 2026 gross sales ₹1.22 Cr vs ₹1.07 Cr on the same orders. Revenue YoY not comparable: Aug 2025 is workbook net (₹56.0 L), Aug 2026 is Shopify gross (₹1.07 Cr) or derived net (₹80.2 L). AOV in the series is net of discounts (Aug: (10,684,206-531,024)/1,960 = ₹5,180).

So whatExtra traffic has hit the same ceiling twice. The cart, not reach, caps orders; Feb tracking break (2.9% coverage) means the cause of Feb's dip cannot be fully separated.

Seasonality

The spring demand dip is seasonal; spend did not flex with it in year two

Oct-Jan vs Mar-Jun, using actual months only. Year one: spend -18.1%, orders -17.0%, CAC flat. Year two: spend +6.0%, orders -11.6%, CAC +19.9%. Only Meta rose in year two (+14.2%); Google fell 17.9%.

The numbers

Y1 (Oct24-Jan25 vs Mar-Jun25): orders 3,399 to 2,822; spend ₹39.9 L to ₹32.7 L; CAC ₹1,174 to ₹1,158; MER 3.34x to 2.94x. Y2 (Oct25-Jan26 vs Mar-Jun26): orders 7,802 to 6,900; spend ₹82.9 L to ₹87.9 L; CAC ₹1,062 to ₹1,274; net revenue ₹3.32 Cr to ₹3.15 Cr; MER 4.01x to 3.58x. Meta ₹61.8 L to ₹70.6 L; Google ₹21.1 L to ₹17.3 L.

So whatPart of the 2026 MER and CAC slide is calendar, part is spend held flat into a weaker quarter. Plan spend by season. Two observations only (two years); Oct-Jan includes festive and wedding peaks.

Correlation, not causation

CAC rose 77% in 11 months and the rise sits almost entirely in Meta spend per order

Blended CAC ₹903 (Aug 2025) to ₹1,597 (Jul 2026). Of the ₹694 rise, ₹570 (82%) is Meta spend per order (₹788 to ₹1,357); Google spend per order moved ₹115 to ₹240 and shows no trend from Oct onward (₹193-362). Meta share of spend fell 93.9% to 78.4% year on year, but Meta spend still doubled (₹90.4 L to ₹1.91 Cr) while Google grew x8.96 (₹5.9 L to ₹52.6 L).

The numbers

Meta Feb to Jun 2026: spend +21.0%, orders -9.6%. Meta spend vs orders r = +0.17 and Google spend vs orders r = +0.71 (Sep 2025-Jul 2026, n=11). Google share of blended spend peaked 31.7% (Nov 2025) and was 15.0% in Jul 2026. Jul 2026 orders (1,538) are the workbook estimate.

So whatThe Google rebuild addressed the smaller cost line. Orders are all-channel, so this is spend allocation per order, not Meta-attributed CAC; Meta results by campaign were never supplied.

Seasonality

March return spike repeats; the Aug 2025 MER peak was flattered by a low return rate

March return rate 35.6% (2025) and 29.4% (2026) against 21.6% and 17.6% for the other 11 months of each year. Net revenue fell 31.4% and 25.4% Feb to Mar while orders fell only 17.7% and 7.9%. February did not repeat (33.5% in 2025, 15.8% in 2026). MER fell in two steps, not gradually: 5.02x (Aug 2025) to 4.05x (Sep), a six-month plateau of 3.82-4.17x (avg 4.04x), then 3.43x (Mar 2026) and 2.68x (Jul 2026). Q3 to Q4 2025 was already -12.4% (4.51x to 3.95x).

The numbers

Order-weighted return rate 22.6% (Aug24-Jul25) to 18.9% (Aug25-Jul26); 19.2% excluding the estimated Jul 2026. Aug 2025 return rate 10.1% is the 24-month low; at 19.83% (average of other year-two months) Aug 2025 MER is ~4.48x, not 5.02x. Mar 2025 net revenue per order ₹2,860 is the 24-month low. MER vs monthly spend r = -0.69 (Aug25-Jul26, n=12) vs -0.15 across 24 months.

So whatReturns lag and post in batches (days 18-31), so month-level MER is noisy. Compare quarters and expect a March return hit in 2027. Workbook return-rate definition is not stated and may differ from Shopify's 19.8% reversals.

Measured pattern

The Search layer got cheaper while its share of Google spend shrank 5x

Search CPA fell ₹117 (Q2 2025) to ₹17 (Q2 2026), CPC ₹5.70 to ₹0.69, CTR 28% to 39% (a brand-term profile). Search cost fell 21.6% (Q3 2025 to Q2 2026) while total Google cost rose x3.94, so Search went from 1.05% to 0.21% of Google cost. Non-search CPA rose ₹969 to ₹1,489 (+53.7%). Search conversions peaked in Q1 2026 (242.6) and fell 32.4% in Q2 (164.0). Q3 2026 CPA ₹165 is the new brand campaign: ₹12,811 for 8.00 conversions (₹1,601); the rest of Search ran ₹3,697 for 91.98 conversions, ₹40.2 (vs ₹17.2 in Q2).

The numbers

Quarterly Search CPA: ₹116.8, 32.1, 34.9, 20.5, 17.2, then 165.1 (Q3 2026, to 17 Aug, partial). Search share of conversions 24.3% (Q3 2025) to 15.3% (Q2 2026) on 0.21% of cost. Overview-export basis, all-action conversions.

So whatBrand demand (~₹64 a day) is a ceiling, not a growth lever; the cheap Search CPA could not absorb budget. My brand campaign explains the Q3 spike, but the inherited layer also worsened 2.3x.

Not a marketing problem

What the data shows ads cannot fix. Saying so early saves budget.

  • Returns: ₹2.41 Cr reversed in 12 months (19.8% of gross); about ₹21.2 L in August vs a ₹10.88 L Google bill. fastrrv3 reverses at 23.8% vs 6.7% native on 75.8% of sales; deity and animal pendants 40-63%, openable kadas 7-15%. RTO vs customer-return split unknown.
  • Cart abandonment: cart-to-order 16.8% to 13.5%; 1,43,983 carts added in a year, 1,23,133 left. At 25% cart-to-order the store does 98.6 orders a day on today's traffic; buying 100 a day costs ₹45-52 L a month.
  • Margin unknown: COGS never supplied. Break-even margin on August's extra spend 33.4% (gross) to 44.6% (net); at ₹29,000/day Stage 5 breaks even at 30%. No campaign can be called profitable.
  • Meta spend: 78.4% of ad spend and 2.26 M sessions, but Aug-Sep 2026 spend never supplied; Oct 2025 monthly sheet vs daily tracker conflicts by ₹55,794.
  • Billing: manual ₹15,000 netbanking top-ups; two blackouts in August (10 Aug about ₹32,000 lost); ₹18,571 left on 1 Sep vs ₹33,000/day.
  • Catalogue and stock: 8,396 SKUs, 478 movers; 2,153 zero-stock variants live; 1,753 variants (34.7%) sold nothing holding 4,574 units; 531 SKUs advertised out of stock; 4.0 months cover.
  • Checkout and attribution: headless checkouts strip the referrer, so only 3.4% of orders carry a source and 93% of Shopify revenue sits under 'None'; Shopify reports 0.01% conversion vs 0.466% true.
  • Orders ceiling: 1,642-2,060 a month since Oct 2025 across 2,36,079-4,20,817 sessions; spring dip repeated in both years (Mar-Jun orders -17.0% in year one, -11.6% in year two).
  • Repeat purchase: 79.6% new customers, month-1 repeat 12.6% decaying to ~1% by month 6; no retention programme proven.
  • Budget brief: stated ₹30 L a month vs ₹20.3 L measured Meta plus Google; if ₹10 L runs elsewhere true CAC is ₹1,700+.
  • Sixteen of the 42 register issues sit outside the ad account (checkout, product mix, finance, data).

Data not available

What would prove more, and does not exist in the sources yet.

  • Gross margin / COGS by category
  • Meta ad spend from Aug 2026, and Meta results (purchases, ROAS) by campaign at any date
  • Monthly returns report: reversal value, reason, SKU and payment mode; RTO vs customer-return split
  • Definition of the workbook 'return %' (may differ from Shopify's 19.8% reversal rate); Jul 2026 net revenue, orders and return rate are estimates from 17 days
  • Shopify gross sales, sessions and AOV by month before Sep 2025; Aug 2025 gross sales, so a like-for-like Aug YoY on revenue is impossible
  • Showroom revenue and walk-in sales; monthly showroom actions before Aug 2026 (only all-time 543)
  • Cause of Google's Jul 2026 conversion rise (+37%) while orders fell; change history for 1-5 Aug 2026
  • Date the checkout-tracking fix went live (coverage 2.9% in Feb, 57.0% in Aug)
  • Go-live date of cart remarketing C04 (timeline marks 31 Aug but August spend is ₹9,999) and the AG02 status (BUILT vs DELIVERED)
  • Outcomes after 3 Sep 2026: tracking repair, call tracking, Stage 3-5 builds, brand campaign fix
  • COD share, remittance days, checkout step events, inventory buy plan Nov-Feb
  • Monthly brand vs non-brand split and quarterly Search layer before 2025-Q2; periods for the 24-month review's AOV and cart-rate pairs
  • Primary evidence for Shopify store work and retention flows (stated by me only)
Level 3

Implementation and data

For the team that has to run it: every action, test, date and dataset.
06

What I actually did

18 actions found in the project evidence, 8 done or live. Grouped by type, each with its source.

Paid media 6

  • Specced

    Stage 3 non-brand Search campaign build sheet: five ad groups, ₹800/day ramping to ₹1,950, 1,246 converting terms

    playbooks/01 s3; mrjewels-seven-weeks index.html s14 (NOT BUILT)
  • Live

    Launched PMax | Store | Bangalore Core for the Indiranagar showroom (1 Aug)

    mrjewels-seven-weeks index.html s02, s04
  • Live

    Wrote three asset groups: Store & Showroom 01, Bridal & Trousseau 02, Festive & Gifting 03

    mrjewels-seven-weeks index.html s04; mrjewels-august-2026-report index.html s11, s14
  • Handed over

    PMax Store negative keywords and surgical negative list (~₹50,000 a month; 67 competitor terms, 32 high-spend terms, off-catalogue tail)

    mrjewels-seven-weeks index.html s09; mrjewels-august-2026-report index.html s10
  • Live

    Demand Gen cart remarketing C04, AG01 Cart 0-7d with image, product and video ads; AG02 Cart 8-30d spec

    mrjewels-seven-weeks index.html s02, s13; mrjewels-august-2026-report index.html s14 (AG02 listed BUILT there, DELIVERED in SW)
  • Handed over

    Merchant Center feed rule to default age_group, gender, color and release ~5,656 throttled products; custom labels 0-3

    mrjewels-seven-weeks index.html s15 item 4; mrjewels-august-2026-report index.html s09

Reporting 3

  • Done

    24-month blended business review (Shopify, Meta, Google) with MER guard rail 3.50x

    mrjewels-seven-weeks index.html s02, s06
  • Done

    All-time Google Ads review (32 tables, 8.6 L terms) and The August Report (16 sections, 22 charts)

    mrjewels-google-ads-case-study docs/index.html; mrjewels-august-2026-report index.html
  • Handed over

    Billing: auto-pay, backup card and three-day balance alert recommended

    mrjewels-august-2026-report index.html s08, s15 item 3

Campaign restructure 2

  • Live

    Rebuilt brand Search with four ad groups (AG01-AG04), AG04 exact plus misspellings built from 168 variant terms

    mrjewels-seven-weeks index.html s02, s08; playbooks/02-brand-adgroup-build-spec.md
  • Handed over

    Search-layer restructure spec: split Brand and Non-Brand campaigns, ₹200/day brand cap, pause Brand Broad 01 (17 Aug)

    playbooks/01-search-layer-restructure.md; mrjewels-seven-weeks index.html s02

Tracking 2

  • Done

    Showroom conversion-goal fix (local actions made primary), 21-25 Aug

    mrjewels-seven-weeks index.html s04
  • Handed over

    Tracking-repair spec: Purchase CAPI out of Conversions, add-to-cart no value, local actions Secondary and scoped, Enhanced Conversions, weekly Google-vs-Shopify check

    mrjewels-seven-weeks index.html s07, s15 item 2

Call tracking 1

  • Specced

    Call-tracking build spec: Exotel numbers on Business Profile, site, Meta and offline; Google forwarding numbers on Search call assets

    mrjewels-seven-weeks index.html s14 'Call tracking', s02 (NOT BUILT)

Dashboard 1

  • Done

    Account command dashboard (HTML + 10-page PDF) and Notion daily ops hub (4 databases)

    mrjewels-seven-weeks index.html s02 (27 Aug)

Strategy 1

  • Specced

    Stage 4 scale plan (50 Google orders a day by November), Stage 5 PMax Core Movers (478 movers, 8 metros) and Volume Core copy (5 asset groups, 75 headlines)

    mrjewels-seven-weeks index.html s14, s02 (NOT BUILT)

Shopify 1

  • Specced

    Shopify store theme and checkout work, prepaid vs cash-on-delivery setup

    earlier portfolio notes MR Jewels 'Shopify store + payments' (in build); no primary report evidence

WhatsApp 1

  • Specced

    Retention: email and WhatsApp lifecycle flows for first-time buyers

    earlier portfolio notes MR Jewels 'Retention: email + WhatsApp' (in build); no primary report evidence
Stated by Danny

In build now. Not yet in the project evidence, so no result is claimed.

  • Call trackingExotel numbers on the Business Profile, site and Meta; Google numbers on Search.
  • Shopify store + paymentsTheme and checkout work, prepaid vs cash-on-delivery setup.
  • Retention: email + WhatsAppLifecycle flows for first-time buyers.
07

Tests and what they taught

Not every test worked. Each one changed the next decision.

  1. Hypothesis

    A bigger brand budget captures more cheap brand demand.

  2. Action

    Raised Search | Brand | Bangalore from ₹2,000 to ₹3,100 a day on 17 Aug, against own spec cap of ₹150-200 (brand demand ~₹64 a day). Campaign sat on Smart Bidding, so ad-group caps did nothing.

  3. Result

    ₹57,249 spent, 23.77 conversions, ₹2,408 per conversion vs ₹3,093 for 84.98 conversions (₹36) on Brand_3june. Broad match bought non-brand clicks.

  4. Learning

    Budget follows measured demand. Check bid strategy before launch. Caught 24 Aug.

  5. Next decision

    Pause Brand Broad 01, one brand campaign with exact CPC cap ₹8-12, broad to phrase, pause 22 dead postcode rows (handed over; outcome not recorded).

  1. Hypothesis

    Google-reported conversion value is reliable enough to scale budgets on.

  2. Action

    Scaled in August: Sales_PMax_CatchAll ₹6,087 to ₹38,806 (+537.5%), Stand_14Jan ₹1,73,483 to ₹2,60,342 (+50.1%); Google billed +143.3%.

  3. Result

    Both campaigns carried defective value (Stand_14Jan ₹12.33 Cr on 1-4 Aug). Shopify: gross sales +21.8%, orders +16.1%, sessions +53.3%, conversion -24.2%; incremental return 2.99x gross, 2.24x net; gross MER 19.61x to 9.82x. S_Tofu scale-up (+268.7%, CPC ₹4.67) was sound.

  4. Learning

    Reconcile Google purchases to Shopify weekly; stop if Google claims more than 60% of orders.

  5. Next decision

    Hold September spend flat at ₹9.0-9.5 L; fix mix and reversals first; scale only above ~45% gross margin (unknown).

  1. Hypothesis

    Purchase tracking is broken: only 33 lifetime add-to-cart events (P0).

  2. Action

    Carried the figure through four documents; planned a Shopify pixel rebuild and a 7 Sep scaling hard-stop on it.

  3. Result

    False alarm: goal cards count primary actions only; add-to-cart fired 26,737 times. Real defects: add-to-cart value bug (GA4 control ₹5,819 per cart), purchase counted twice (₹4.74 Cr claimed vs ₹2.98 Cr real, 1.6x), no primary action on two goals (35% of 1-2 Sep conversions not sales).

  4. Learning

    Read conversion actions, not goal tiles. Cost: a week of gated decisions. Caught 2 Sep.

  5. Next decision

    Purchase CAPI out of Conversions, add-to-cart no value, one live test order, 7-day reconciliation, then value net of reversals. Do not promote add-to-cart to primary.

  1. Hypothesis

    Cart remarketing recovers abandoned carts.

  2. Action

    Launched Demand Gen C04, cart 0-7 day audience, three ads, with no stop-or-continue date; second ad group specced on the same base.

  3. Result

    By 2 Sep: ₹11,971, 3,118 clicks, 0 purchases; 92 view-through conversions argued for patience.

  4. Learning

    Set the stop date before launch. Cost ₹11,971.

  5. Next decision

    Two more weeks with corrected cart audience and a conversion action attached; all-conversions check pending (open at close).

  1. Hypothesis

    No local action was primary, so store bidding was blind.

  2. Action

    Set local actions as the primary conversion for PMax Store, 21-25 Aug.

  3. Result

    CPC ₹38.60 to ₹13.13 (-66%), CTR 0.69% to 2.40%, clicks 122 to 1,241, cost per local action ₹38.24.

  4. Learning

    Highest-value single edit of the engagement. The 34.06% rate counts directions and calls, not sales.

  5. Next decision

    Fund Bridal and Festive asset groups before 6 Nov (Dhanteras, Diwali); keep local actions Secondary outside PMax Store.

  1. Hypothesis

    A 17-day July can be scaled linearly to a month.

  2. Action

    Blended review scaled July by 31/17.

  3. Result

    July MER published at 3.20x, restated to 2.68x the same day: days 1-17 carry 66.1% of a month's net revenue because returns post on days 18-31.

  4. Learning

    Never linear-scale a month. The Jul 2026 workbook point remains an estimate.

  5. Next decision

    Use full months only; request monthly returns report.

08

Timeline: interventions and events

My actions sit alongside what the client, the previous team and the market did. Nothing here claims a cause.

  1. Previous team

    Google Ads account starts spending; run by a previous team for 25 months

    mrjewels-seven-weeks index.html s01
  2. Client

    Blended workbook begins: net revenue ₹10.3 L, Meta ₹1.5 L, MER 6.14x

    client monthly history (client MER workbook, seven-weeks js BL)
  3. Previous team

    Google account scales hard from this month (monthly spend ₹31k to ₹4.6 L by Oct 2025)

    mrjewels-google-ads-case-study docs/index.html s03; hist google_cost_inr
  4. Market

    Blended MER peak 5.02x with return rate at 24-month low 10.1%

    client monthly history (client MER workbook, seven-weeks js BL)
  5. Previous team

    Google share of blended ad spend peaks at 31.7%

    client monthly history (client MER workbook, seven-weeks js BL) (my calc)
  6. Client

    Gross sales peak ₹1.22 Cr; checkout tracking breaks (2.9% coverage) and Google impressions peak at 97.2 L

    mrjewels-august-2026-report index.html s07; hist google_impressions
  7. Market

    Return rate 29.4% and net revenue -25.4% on February

    client monthly history (client MER workbook, seven-weeks js BL) (my calc)
  8. Previous team

    Blended MER 2.68x (restated), worst month in two years; Google conversions +37% while orders -5%

    mrjewels-seven-weeks index.html s06; hist
  9. Danny

    Pitch: account review and 12-month media plan

    mrjewels-seven-weeks index.html timeline
  10. DannyI join

    PMax Store and brand Search live; Stand_14Jan books ₹12.33 Cr phantom value 1-4 Aug

    mrjewels-seven-weeks index.html s02, s07
  11. Client

    Balance exhausted; spend ₹533 for the day

    mrjewels-august-2026-report index.html s01
  12. Danny

    All-time Google Ads review; AG04 live; brand budget raised ₹2,000 to ₹3,100/day

    mrjewels-seven-weeks index.html s02, s13
  13. Danny

    Stage 3 non-brand Search spec (₹800 to ₹1,950/day) delivered, not built

    mrjewels-seven-weeks index.html s02
  14. Danny

    24-month blended review; Bridal asset group built; margin and Meta spend asked for

    mrjewels-seven-weeks index.html s02, s15
  15. Danny

    Bid-strategy pre-flight catches brand budget error; AG05 and store negatives delivered

    mrjewels-seven-weeks index.html s13
  16. Danny

    Showroom conversion-goal fix complete (21-25 Aug); Festive & Gifting asset group live

    mrjewels-seven-weeks index.html s02, s04
  17. Danny

    Account command dashboard (HTML + 10-page PDF) and Notion ops hub delivered

    mrjewels-seven-weeks index.html s02
  18. Danny

    Call-tracking spec (Exotel), Road to 100 orders a day, Stage 4 remarketing spec

    mrjewels-seven-weeks index.html s02, s14
  19. Danny

    Cart remarketing C04 shown live; balance exhausted again (₹18,571 left 1 Sep)

    mrjewels-seven-weeks index.html s02
  20. Danny

    The August Report published; false P0 and sample-misread corrections found

    mrjewels-seven-weeks index.html s02, s13
  21. Danny

    Tracking repair spec and Stage 5 Core Movers spec handed over

    mrjewels-seven-weeks index.html s02, s07
  22. Danny

    Last build: Volume Core asset-group copy (5 groups); execution after this date not recorded

    mrjewels-seven-weeks index.html s02
09

The data, month by month

Strip every Google conversion number. Use only Shopify revenue and the billing screen. See which story survives.

Two years before I arrived: revenue nearly tripled

Net revenue by month, ₹ L, client workbook. Year one ₹3.17 Cr → year two ₹9.21 Cr (Jul 2026 estimated in the workbook). Another team's work. Measured

Paid mix, 24 months: Meta carried the spend

MetaGoogle

Monthly ad spend, ₹ L, client workbook. Google grew from under ₹1 L to ₹3–7 L a month in year two. Measured

…while each ad rupee returned less

Blended return on Meta + Google spend. 5.02× in Aug 2025 (flattered by a 10.1% return month; ~4.48× at a normal rate) → 2.68× in Jul 2026. Calculated

August: ₹1.07 Cr, above the 12-month average

Shopify gross sales, ₹ L. Purple: the month I took over. The 11 months before averaged ₹1.01 Cr. Measured

The ₹12.56 Cr that never existed

One Shopping campaign booked ₹12.33 Cr in four days on ₹53,896 of spend. Measured

81% of the money sat in the costliest format

Share of spendShare of conversions

All-time, Jul 2024 – Jul 2026, before my work. Search ran at ₹73 per conversion on 1% of budget. Calculated

One goal fix: showroom CPC −66%

No local action was set as primary, so bidding bought anonymous clicks. Measured

Spend +143%. Orders +16%.

July → August. New traffic converted a quarter worse: wider, not better. Measured

Meenakari chandbali earrings · mrjewels.in
Gold bead chain · mrjewels.in

Evidence viewer: the client reports

Engagement reportPreview the reportOpen full screen →
August reportPreview the reportOpen full screen →
All-time account reviewPreview the reportOpen full screen →
10

What I'd do next

SourcesClient MER workbook Aug 2024–Jul 2026 (24 months, ₹12.4 Cr net revenue, 29,247 orders), Shopify Sep 2025–Aug 2026, Google Ads all-time and August exports (15.8 lakh search-term rows), Merchant Center, Shopify sales, orders, sessions, returns, landing pages and customers, client MER workbook, billing 2024–2026.Evidence labels: Measured, Calculated, Estimated, Assumption, Data not available.
← Previous caseTONI&GUY Bangalore+14.7% · salon revenue vs same months of 2025
Next case →Elites Dental Care₹3,140 · per booking, before the rebuild

Work with me

Seen a leak in your own numbers?

Send me the symptom. I'll tell you which case is closest and what I'd check first. No deck, no retainer pitch.

danieldeepak3519@gmail.comLinkedInFree 12-point check