Agency Insights
Before Nurotech takes on a pay per click service in India engagement, we run one audit. It takes 90 minutes on a typical account. In 2025 we ran it on 41 accounts before onboarding. The pattern held in 34 of them. At least four of the eleven leaks below were live in each one. The average account was quietly giving back ₹18,000 to ₹65,000 a month to settings nobody had looked at since the account was built. None of it requires a new agency or a bigger budget. It requires an hour with the right screens open.
This is that audit, leak by leak. Each section names the exact screen where the leak hides, the rupee range we've seen it cost, and the fix. Run it on your own account before you hire anyone — including us. If you find three or more of these live, that is the actual reason your cost per lead climbed. It was never "the algorithm changed."
Why We Audit the Account Before We Touch the Budget
Most pay per click agency pitches in India open with a media plan. We open with an audit instead. According to our own 2025 review of onboarded accounts, the single biggest lever in the first 60 days was never the bid strategy. It was turning off spend that was never supposed to be on. A founder who inherited an agency-run account almost never has this data. The previous vendor's monthly report shows clicks and conversions. It does not show where the clicks came from within the network. That gap is where the leaks below survive for years.
Leak 1 — Search Partners and Display Expansion Left On by Default
Path: Campaigns > Settings > Networks. Google Ads ships new Search campaigns with "Include Google Search Partners" checked. On some account types, display expansion is checked too. Search Partners sends your ad to third-party search boxes and portal sites you cannot see individually. You only see the aggregate in the Networks report, under Campaigns > Reports > Predefined reports > Basic > Network.
On accounts spending ₹2,00,000 to ₹5,00,000 a month, we routinely find 12% to 28% of Search spend sitting on Search Partners. The conversion rate there runs roughly half of Google Search proper. In our experience the fix is a five-minute checkbox change. It alone recovers ₹15,000 to ₹40,000 a month on a mid-size account, without touching bids at all.
Fix: open the Network report for the last 90 days. Compare Search Partners CPA to Google Search CPA. Uncheck Search Partners on any campaign where the partner CPA runs above 1.5x the Search CPA.
Leak 2 — Broad Match Plus Smart Bidding on a Thin Conversion History
Broad match paired with Target CPA or Maximize Conversions works when the campaign has enough conversion volume to learn from. Google's own documentation states the bidding model needs a steady flow of conversions to calibrate. In practice, we do not trust it below roughly 30 conversions in a rolling 30-day window. Below that line, broad match plus Smart Bidding on a new or small account spends the learning-phase budget on queries that only vaguely relate to the product. The algorithm is guessing, not learning.
We tracked this across 14 new accounts we onboarded between January and June 2025. The ones we left on broad match plus Target CPA in month one averaged a 40% higher cost per lead than the same accounts on Phrase match with manual or Enhanced CPC bidding. The gap traced entirely to match-type breadth on thin data, nothing else.
Fix: if your account runs under 30 conversions a month, drop to Phrase match and rebuild volume first. Only widen to broad match once the account has a real conversion history to bid against.
Leak 3 — Brand-Term Cannibalisation
Path: Campaigns > Search terms, filtered to your own brand and close misspellings. If your organic listing already ranks #1 for your brand name, you are paying ₹8 to ₹35 a click — depending on the vertical — for a click Google would have handed you for free. This is the most common single leak we find. It shows up in roughly 60% of the accounts we've audited.
The catch: some brand spend is genuinely defensive, if a competitor is bidding on your name. Per the Google Ads auction insights report, a competitor holding over 10% impression share on your brand terms changes the calculus. Some brand spend is worth keeping in that case. Absent a competitor, it is close to pure waste.
Fix: pull Auction Insights for your brand campaign. No competitor above 10% impression share → pause or cap brand spend at under 5% of budget. Competitor present → keep a lean brand campaign, but cap the daily budget rather than letting Smart Bidding chase every click.
On the compliance side, the Ministry of Electronics and Information Technology's IT Rules guidance is worth a read if your ads or landing pages make any data-collection or consent claims. Several accounts we've audited had a form collecting more data than the on-page privacy language disclosed.
Leak 4 — Performance Max Eating Search
Performance Max campaigns pool Search, Display, YouTube, Discover and Gmail inventory into one black-box campaign. When a Search campaign and a PMax campaign run in the same account without exclusions, PMax will win the auction on your own brand and category terms. Google's internal auction favours the campaign type carrying the most signal, and PMax carries more by design. The Channel report — Insights > Performance breakdown > Channel — is the only place this shows up clearly. Most advertisers never open it.
Our data set from Q1 2025, across nine PMax-plus-Search accounts, found that PMax was pulling 35% to 55% of what should have been brand Search traffic. The cost per click there ran materially higher, because PMax was bidding against the account's own Search campaign.
Fix: add your brand terms as a negative keyword list at the account level, so PMax cannot bid on them. Check the Channel report monthly — this setting silently re-breaks after almost any campaign edit.
Leak 5 — Empty or Stale Negative Keyword Lists
Path: Campaigns > Search terms, sorted by cost, no conversions, descending. Our audit found zero negative keywords added in the trailing six months in 22 of the 41 accounts we reviewed in 2025. Every irrelevant query the algorithm decided to try had stayed live indefinitely.
A "pay per click company in india" search for a B2B SaaS client, for example, regularly pulls in "pay per click jobs," "pay per click salary" and "what is ppc." None of these convert. All of them cost real money, at ₹8 to ₹20 a click. On accounts we've fixed this for, the search-terms cleanup alone recovered 8% to 15% of monthly spend.
Fix: this needs a weekly cadence, not a one-time pass. Every week, sort search terms by spend with zero conversions over the last 30 days. Add anything irrelevant to a shared negative list across all campaigns.
Leak 6 — Conversion Double-Counting Between Google Ads and GA4
Path: Goals > Conversions > Summary, cross-checked against GA4's Google Ads linked-account conversions. When a GA4 property links to Google Ads, and both the native Google Ads tag and a GA4 conversion event get imported as separate conversion actions, the same lead can be counted twice inside Google Ads' own reporting. That inflates the conversion count the bidding algorithm optimises against.
We measured this on six accounts in early 2025. Marketing was reporting a 4.2x ROAS to leadership. The true, de-duplicated number, once we mapped which conversion actions were imported versus native, came out to 2.6x. The bidding algorithm had been optimising toward the inflated number the entire time. That pushes spend toward audiences and placements that only look like they convert.
Fix: in Conversions > Summary, check the "Source" column for every active conversion action. If the same event has both a native Google Ads tag and a GA4 import, mark one as "Not used in Targets" so the bidding model isn't double-fed.
Leak 7 — Conversion Actions That Count Everything as a Lead
This is the leak we see most often on accounts run by agencies billing on cost per lead. The conversion action is set to "Page view — thank-you page," rather than an actual form submission with a validated phone number. Or worse, a phone-call conversion counts any call over 10 seconds, including wrong numbers and hang-ups.
Our team measured actual sales-qualified lead rate against the reported conversion count on four client accounts we took over in 2025. The gap ranged from 20% to 48%. Up to half of "conversions" a previous agency reported were never a real inquiry.
Fix: audit every conversion action's trigger condition under Goals > Conversions. Replace page-view-based conversions with form-submit events tied to CRM-validated leads. Set phone-call conversions to a minimum duration of 60 seconds, not 10.
Leak 8 — Geo Targeting Set to "Presence or Interest"
Path: Settings > Locations > Location options. The default targeting setting is "Presence or interest." It serves your ad to anyone anywhere in the world who has merely searched about your target city — not people actually located there. For a business that only serves Delhi NCR, this means paying for clicks from someone in another state or country who typed "gyms in Delhi" out of curiosity.
Across the accounts we've audited that serve a specific city or region, roughly one in three had this default left untouched. The "interest" share of clicks ran 10% to 25% of total volume, with a conversion rate near zero, since the searcher was never local to begin with.
Fix: switch to "Presence: People in or regularly in your targeted locations" under location options. It is a single dropdown change with no downside for a locally-served business.
Leak 9 — Landing-Page Mismatch and the Quality Score Cost
Path: Keywords > Status column, expand for Quality Score components. When the ad headline promises one thing and the landing page says something else — "Free Consultation" in the ad, no mention of "free" anywhere on the page — Google's Landing Page Experience component of Quality Score drops. A lower Quality Score raises the actual CPC you pay for the same ad rank, sometimes by 20% to 40%.
Seven of ten accounts we onboarded in 2025 had at least one ad group where the promised offer in the headline never appeared on the destination page above the fold.
Fix: for every ad group, open the top three ads and the landing page side by side. The specific promise in the headline — price, offer, timeframe — needs to appear in the page's own H1 or the first screen, not buried lower.
Leak 10 — No Account-Level Conversion Value
Path: Tools > Conversions > column "Value". Without a value assigned per conversion action, Smart Bidding optimises purely for conversion count. A ₹500 low-quality lead and a ₹50,000 qualified enterprise lead are worth exactly the same to the algorithm. It will happily chase more of the cheap kind, because they are easier to get.
On one D2C account we audited, we assigned realistic conversion values — even estimated ones, based on average order value by product category — and switched the bid strategy to Maximize Conversion Value. Spend shifted toward the higher-value product lines. Blended ROAS improved by roughly 30% over the following quarter, with no change in total budget.
Fix: assign a static or dynamic value to every conversion action, based on average deal size or order value. Then move the bid strategy from Maximize Conversions to Maximize Conversion Value.
Leak 11 — Attribution Window and Data-Driven Attribution Set Too Early
Path: Goals > Conversions > select action > Attribution settings. Data-driven attribution needs a meaningful volume of conversion and click data to build a reliable model. Google's stated minimum is roughly 300 conversions in 30 days across the account, per the Google Ads attribution documentation. Accounts switched to DDA before hitting that volume get a model trained on too little data. It can misassign credit across keywords and destabilise Smart Bidding decisions.
We recommend last-click or a fixed attribution model until an account genuinely clears that volume threshold. Then a controlled switch, with two weeks of before/after comparison, before trusting the new numbers.
Fix: check trailing-30-day conversion volume before enabling Data-Driven Attribution. Below roughly 300 conversions account-wide, stay on last-click.
If your account also handles telemarketing follow-up on form leads, the Telecom Regulatory Authority of India's guidelines on commercial communication are the other compliance layer worth checking. Several leads we've audited turned out to be flagged do-not-disturb numbers, being called anyway.
A Worked Example: One Account, Three Leaks, ₹42,000 Recovered
We onboarded a Delhi-based B2B services client in March 2025, spending ₹3,80,000 a month across Search and PMax. The prior agency's monthly report showed a healthy 3.1x ROAS. Our audit found three leaks live at once: Search Partners uncapped, brand terms being eaten by PMax, and a negative keyword list last touched fourteen months earlier.
We fixed all three in the first two weeks, in this order: brand negatives at the PMax level first, Search Partners second, search-terms cleanup third. We tracked cost per qualified lead weekly for eight weeks after. Week one showed almost no change — the fixes need a few days to clear the reporting lag. By week three, cost per lead had dropped from ₹1,850 to ₹1,290. By week eight, it settled at ₹1,110, a 40% reduction with the same monthly budget.
The true ROAS, once we corrected the double-counted GA4 conversions we also found in month two, was 2.4x at the start — not the reported 3.1x. By month three, on the same de-duplicated basis, it reached 3.6x. That is the recovery curve in practice, not theory: ₹42,000 a month in spend that used to leave the account through three settings, now going to keywords that actually convert.
Which Leaks Hit Hardest, by Business Type
Not every leak costs the same across every account type. Our data set splits fairly cleanly by business model.
D2C and e-commerce accounts lose the most to Leak 10 — no conversion value — because product margins vary widely, and Smart Bidding without a value signal treats a ₹300 order the same as a ₹15,000 one. We've seen this leak alone account for over half the wasted spend on the D2C accounts we've audited.
B2B service accounts, the kind selling a pay per click service in india or similar consulting work, lose the most to Leak 7 — inflated conversion actions — and Leak 3, brand cannibalisation. Long sales cycles make it tempting to count every form fill as a win, which then feeds a bidding algorithm optimising for volume over quality.
Local service businesses — clinics, gyms, home services — lose the most to Leak 8, geo targeting set to "Presence or interest," because their entire value proposition depends on the searcher being physically nearby. A yoga studio in Gurgaon paying for interest-based clicks from Mumbai is paying for a click that can never convert into a walk-in client.
Knowing which category your business falls into before you start the audit saves time — start with the leak most likely to be live for your model, then work through the rest.
What a Clean Account Looks Like: The 60-90 Day Recovery Curve
Fixing all eleven leaks in one sitting is tempting, and wrong. Changing bidding, attribution, conversion actions and targeting all at once makes it impossible to tell which fix moved the number. On the accounts we've taken through this audit, the realistic sequence runs in three phases.
Days 1-14: the free fixes — Search Partners, geo targeting, brand negatives at the PMax level, obvious search-term negatives. These carry no learning-phase cost. They typically recover 15% to 25% of wasted spend within two weeks.
Days 15-45: conversion-action cleanup and value assignment. This step often makes the reported conversion count drop before it improves, because you're removing the inflated page-view and short-call conversions. Expect the dashboard to look worse for one to two weeks before the true, de-duplicated numbers stabilise.
Days 46-90: bid-strategy and attribution changes, once the conversion data feeding them is honest. This is where cost per qualified lead typically falls 25% to 45% from the account's starting point, based on the pattern across dozens of accounts run through this same sequence.
None of this requires a bigger budget. It requires stopping the leaks before asking the algorithm to optimise around them. Nurotech's performance marketing team runs this exact sequence for every new Google Ads client. This article walks through the same audit, unedited.
If you'd rather have us run this audit on your account directly, our performance marketing service starts every new engagement with it, at no extra charge. See what else is included on our pricing page, or explore our full service list. If your paid budget is split across Meta as well, our companion piece on how to evaluate a Meta ads agency in India covers the same ground for that channel.
Frequently Asked Questions
The short answers below cover the questions we get asked most before a Google Ads audit engagement. Fuller detail on scope and pricing is on our performance marketing page.
How Long Does a Full Google Ads Audit Take?
Our own audit runs 90 minutes for an account of typical size — under ₹5,00,000 in monthly spend, five to ten campaigns. Larger, multi-market accounts with 20+ campaigns can take half a day.
Who Should Own the Google Ads Account — the Agency or the Client?
We set up every client account under the client's own Google Ads MCC login, never the agency's. It costs nothing extra. It also means you keep full data history and access if you ever change agencies.
What Is a Realistic Minimum Budget for Google Ads in India?
Below roughly ₹50,000 a month in ad spend, most accounts cannot generate the 30 conversions a month that Smart Bidding needs to calibrate. Results tend to be inconsistent regardless of who manages the account.
Can I Run This Audit Myself Without an Agency?
Yes — that is the point of listing the exact screen path for each leak above. Anyone with admin access to their own account can run through all eleven checks in under two hours.
Does Fixing These Leaks Guarantee Lower Cost Per Lead?
No single fix guarantees an outcome, and we don't promise one. What we can say is that across the accounts we've audited, fixing three or more of these leaks has consistently reduced wasted spend. That is a different claim than promising a specific result.
How Often Should This Audit Be Repeated?
We re-run the full eleven-point check quarterly on every account we manage, and the lighter weekly negative-keyword pass never stops. Google ships new default settings and new campaign types often enough that a leak fixed in January can quietly reappear by June — Leak 4, PMax eating Search, is the one we see resurface most, usually after a well-meaning account edit resets the brand exclusion list. Treat this as a recurring maintenance task, not a one-time cleanup.
