Digital Marketing Agency in Pune: When to Move Your Lead-Generation Google Ads From Manual Bidding to Smart Bidding (and When to Stay Put)
Move a lead-generation Google Ads campaign to Smart Bidding when it records enough real conversions for the system to learn from, when those conversions actually represent leads you want, and when you can leave the campaign alone long enough to judge the result. Google's own guidance is to evaluate performance over the last 30 days with at least 30 conversions. If a campaign produces five form fills a month, or counts junk enquiries as conversions, switching on Target CPA will not rescue it; it will optimise confidently towards the wrong thing. A digital marketing agency in Pune should be able to tell you which side of that line your account is on, show you the numbers, and explain what it will do if the switch goes badly.
This article is for business owners and marketing managers who pay for Google Ads, or are deciding whether to, and want to judge an agency's bidding advice. It covers what Smart Bidding is, the four strategies and their plain-English meaning, what "enough data" means, why lead quality comes before the bidding choice, a step-by-step migration, how to set a target, how to read results fairly, common failure patterns and questions to ask. Facts come from Google Ads Help pages read in October 2026: the Smart Bidding overview, the Target CPA article and the article on learning-period duration. Google changes interface labels and policies often, so confirm the current version of any page you rely on. Where I describe general practice rather than a documented rule, I say so. The worked example is hypothetical, with illustrative numbers, and no real client is described. This article does not cover payments, shipping or tax.
The short answer a digital marketing agency in Pune should give you
Manual bidding means you (or your agency) set the maximum cost per click for keywords and adjust it by hand. Smart Bidding means Google's system sets the bid for each individual auction, using signals such as device, location, time and query, to pursue a goal you choose. Google's Smart Bidding overview describes it as using Google AI to optimise for conversions or conversion value in every auction.
For many lead-generation accounts Smart Bidding eventually wins, because the system can weigh far more signals than a person can. But three conditions should hold before you switch:
- Conversion tracking is accurate. Google's page lists conversion tracking as a requirement. If it counts duplicate form submissions, button clicks that are not leads or calls under a few seconds, the system learns to find more of those.
- There is enough volume. The Smart Bidding overview recommends measuring over a period with at least 30 conversions. Its Target ROAS guidance mentions 50. Per-strategy minimums vary by campaign type, so check the strategy's own page.
- A conversion means a lead you want. If half your "conversions" are job-seekers, wrong-city enquiries or spam, the bidding will optimise for them.
If one of these fails, fix it first. Often the right answer to "should we switch to Smart Bidding?" is "not yet, here is what we do for the next eight weeks".
What the four Smart Bidding strategies mean
Google's overview groups four strategies by business goal. In plain English:
| Strategy | What it tries to do | Typical fit |
|---|---|---|
| Maximize conversions | Get as many conversions as possible within your budget | Lead generation with a fixed monthly budget and no strict cost limit |
| Target CPA | Get as many conversions as possible at an average cost per conversion you set | Lead generation once you know what a lead can cost |
| Maximize conversion value | Get as much total conversion value as possible within budget | Accounts where leads carry different values (for example, by service) |
| Target ROAS | Get as much conversion value as possible at a return on ad spend you set | Stores or lead accounts with reliable value data |
Two details from the pages I read are worth knowing. First, the Target CPA article says that if Maximize conversions is available for your campaign type, Google recommends using it rather than Target CPA. Second, the overview notes that from June 2026 the labels "Maximize conversions with a Target CPA" and "Maximize conversion value with a Target ROAS" become simply "Target CPA" and "Target ROAS", with the underlying behaviour unchanged. If your agency's reports use older names, that is why.
For a lead-generation business with no ecommerce transaction values, the realistic choice is between Maximize conversions and Target CPA. The value-based strategies only make sense if you can send Google meaningful value numbers, which usually means connecting your CRM, a topic covered in our article on lead quality and offline conversions.
How much data is enough?
Google does not publish one magic number for every campaign. What the pages I read do say:
- Evaluate performance over the last 30 days and at least 30 conversions (Target CPA article, repeated in the overview).
- Some strategies work best with a minimum of historical conversion data, which depends on campaign type; the overview points to each strategy's own article for details.
- Target CPA can start with no conversion history, and Google describes it as effective for campaigns of all sizes, though measuring it reliably still needs the 30-conversion window.
Translate that into a rough rule for a small Pune business: if the campaign delivers fewer than about 30 genuine conversions in a month, the bidding system will have thin data and your own evaluation will be noisy. That is a rule of thumb derived from Google's evaluation guidance, not a published eligibility requirement. Low volume does not forbid Smart Bidding, but it does mean results are harder to read, so expect to wait longer before judging.
Options when volume is low:
- Consolidate campaigns. Ten tiny campaigns each starved of data usually perform worse than two or three with enough conversions each.
- Use a more frequent conversion signal. A "qualified call" or "form submit" counted alongside a deeper event, with the deeper event marked as the primary goal, gives the system more to learn from. Choose carefully: a shallow signal such as page scrolls is not a lead.
- Broaden carefully only if the extra traffic is relevant; more clicks that never convert do not help the learning.
- Stay on manual bidding or Maximize clicks with strict negative keywords for now. This is legitimate, as long as you are clear about why.
Lead quality comes before the bidding strategy
The most expensive mistake in lead generation is optimising a campaign for the cheapest form fill. Smart Bidding is a very effective machine for finding exactly the kind of conversion you have defined. If the form is easy and the offer is vague, it will find many people who fill forms and never buy.
Before changing bidding, check:
- What counts as a conversion? Is it a form submission, a call over a certain length, a WhatsApp click? Are duplicates removed?
- What happens to leads afterwards? If your sales team marks leads as junk, qualified or won, that data can be imported back into Google Ads so the system learns from outcomes rather than form fills. See our guide on Enhanced Conversions and Consent Mode for why accounts lose lead data and how to recover it.
- Are search terms relevant? Smart Bidding sits on top of your targeting. If irrelevant queries are getting through, tighten the account with negative keywords first; our article on search terms and negative keywords explains the routine.
- Is the landing page converting? A page that loses visitors raises your cost per lead whatever the bidding does.
A reasonable agency will ask these questions before proposing a switch. One that proposes Target CPA in the first week, without looking at them, is selling a setting rather than a result.
A staged migration plan
Here is a sequence a careful agency might follow. The stages are general practice, not a Google-mandated process.
1. Audit the measurement (week 0). Confirm what each conversion action counts, remove duplicates, and compare Google Ads conversions with the enquiries in your CRM or inbox for a sample week. If they differ widely, resolve that first.
2. Clean the account (weeks 0 to 2). Add negative keywords from search term reports, pause keywords with spend and no relevance, and fix the landing page issues you can fix quickly.
3. Build a baseline (weeks 0 to 4). Record, for the last 30 days: spend, clicks, conversions, cost per conversion, and the share of conversions that your sales team judged qualified. This is the yardstick for everything after.
4. Choose the first strategy. For most lead-generation accounts, Maximize conversions is the simpler first step because it needs no target. Once it has run steadily and you know the typical cost per conversion, consider Target CPA. Google's recommendation to prefer Maximize conversions where available fits this approach.
5. Test rather than flip. Google's Target CPA article describes a "Save as experiment" option to compare new settings against the original campaign instead of saving the change immediately. Using an experiment, or a campaign experiment run for several weeks against the existing setup, lets you compare like with like. If experiments are not available for your campaign type, change one campaign and leave a comparable one alone, accepting that the comparison is rougher.
6. Hold still. Avoid changing the budget, keywords, ads and target all at once. Each is a separate change, and stacking them makes the results impossible to interpret.
7. Evaluate against the baseline. After the evaluation window, compare cost per qualified lead, not just cost per conversion.
8. Adjust gradually. If you set a target, move it in small steps.
Setting a target without guessing
If you use Target CPA, the number matters. The Target CPA article says that with historical data, Google Ads recommends a target based on the average CPA over the prior 30 days, adjusted for conversion delays; without history, it uses your business goals. You can accept the recommendation or enter your own.
Be careful in two directions:
- Too low. The article warns that setting the target too low can cause the system to skip clicks that might have converted, which means fewer total conversions. If your lead volume collapses after a change, check the target first.
- Unrelated to your economics. A target should come from what a lead is worth to you. A simple calculation: if one in ten leads becomes a customer and a customer is worth a margin of ₹X to you, a lead is worth roughly X divided by ten. Compare that with the current cost per lead. These are your numbers, and I have deliberately not invented typical figures for Pune industries, which vary hugely.
Another detail from the same article: compare the cost you achieve with the average target CPA shown in the account, not simply the number you typed, because the average reflects device adjustments, ad group targets and past changes. It also notes that actual CPA can differ from the target for reasons outside Google's control, including website changes, ad changes and auction competition.
A further notice on the same page says that from August 17, 2026, budget-limited campaigns using Target CPA or Target ROAS may see temporary performance and traffic fluctuations. If your campaigns are capped by budget, ask your agency whether this affects you and how it will tell the difference between that and a real problem.
The learning period, and why patience is part of the strategy
After a change, the bidding system calibrates. Google's help article on the duration of the learning period says there is no fixed number of days: duration is mainly affected by the number of conversions, the length of conversion cycles and the bid strategy type. It adds that it can take up to around 50 conversion events or three conversion cycles for a strategy to calibrate to a new objective, though it can be faster with more data. The status column in Google Ads can show "Learning" with a reason: a new strategy, a setting change, or a composition change.
What that means for a business owner:
- Do not judge after three days. Older third-party blog posts quote a seven-day learning period; Google's current documentation does not use a fixed number. A business with a long sales cycle, where a form fill takes weeks to turn into a qualified lead, will need longer.
- Make fewer changes. A target change counts as a setting change and can put the strategy back into learning. Frequent tweaks mean the system is always recalibrating.
- Expect some volatility while the status shows learning, and agree in advance with your agency how long you will wait and what would make you intervene sooner (for example, spend running well beyond budget or lead volume dropping to nothing).
Google also notes that the algorithms continue to learn even when the status no longer says "Learning". A stable account is one where changes are deliberate and spaced out.
A worked example (hypothetical)
Take a Pune physiotherapy clinic that runs Search ads for "physiotherapy near me" type queries. The numbers below are illustrative, made up to show the method; they are not from a real client or an industry benchmark.
Starting point: manual bidding, a monthly budget, and 14 form fills in the last 30 days. Of the 14, the front desk marked 6 as real appointment requests.
Diagnosis: 14 conversions is under the 30-conversion evaluation guidance, and fewer than half are qualified. A switch to Target CPA now would be optimising on thin and partly unreliable data.
Fix first:
- Remove a duplicate conversion that counted both the button click and the thank-you page.
- Add the clinic's calls from ads of 60 seconds or more as a second conversion, and import "appointment booked" from the clinic's records weekly.
- Add negative keywords for job and course searches found in the search term report.
- Rewrite the landing page form to ask for the condition and preferred time, which discourages casual enquiries.
Eight weeks later: with the duplicate removed and the new signals, the account records enough genuine conversions to evaluate over a 30-day window. The agency moves one campaign to Maximize conversions and keeps a comparison campaign on the old setup, then reviews cost per booked appointment, not cost per form fill.
Decision rule agreed in writing: if cost per booked appointment is not better after the evaluation window, revert; if it is, consider Target CPA with a target set at the recent average, and move in small steps.
The point is the sequence: measurement, then cleanup, then a controlled test with a stated success measure.
Common failure patterns
- Switching on Target CPA with a target set from wishful thinking. The system skips clicks and volume falls.
- Counting the wrong conversions. Page views, scrolls or button clicks treated as leads inflate numbers and pollute learning.
- Judging too early. A reversion on day four restarts the cycle and teaches nothing.
- Changing everything at once. New ads, new keywords, new budget and new bidding in the same week; no one can tell what worked.
- Ignoring seasonality. A festival-season spike or a dull month changes conversion rates regardless of bidding.
- Reporting on the platform's own conversion counts only. Reconcile with the CRM.
- Starving the budget. If the daily budget is tiny relative to the target, the system cannot buy enough auctions to learn.
- No exit plan. Always agree beforehand when you would revert.
Questions to ask a digital marketing agency in Pune
- How many genuine conversions did the campaign record in the last 30 days, and how many were qualified?
- What exactly counts as a conversion, and have duplicates been removed?
- Which strategy do you propose, and why that rather than Maximize conversions or staying manual?
- How will you test it, by experiment or a comparison campaign, and for how long?
- How will you decide a target CPA, and what will you do if volume drops?
- Which measure decides success: cost per conversion or cost per qualified lead?
- What changes will you avoid during the evaluation window?
- How do you handle the learning status and temporary fluctuations?
- Can I see conversion numbers reconciled with my own records?
- What is the revert plan, and who decides?
Our article on what a monthly agency retainer should deliver helps with framing the reporting side of these questions.
Working with Govindani Infotech
Govindani Infotech is a web design and development company at 2nd Floor, Landmark Plaza, 206 Satara Road, Parvati Paytha, Pune 411009, and we also run paid search and social campaigns for clients. If you would like a second opinion on your current Google Ads bidding, we can look at what the account counts as a conversion, how many it has recorded recently and whether Smart Bidding is likely to help, then tell you plainly if staying put is the better choice.
Our Google Ads service page explains how we set up and manage campaigns, and our website creation work is relevant when the real problem is the landing page. To talk it over, use the contact page or message us on WhatsApp.
Frequently Asked Questions
When should a small business switch from manual bidding to Smart Bidding?
When tracking is accurate, conversions represent leads you actually want, and the campaign records enough of them to learn from. Google recommends evaluating performance over a period with at least 30 conversions, so a campaign with only a handful each month may need consolidating or better conversion signals first.
What is the difference between Maximize conversions and Target CPA?
Maximize conversions tries to get as many conversions as possible within your budget, with no cost target. Target CPA tries to get as many as possible at an average cost per conversion you set. Google's Target CPA article recommends using Maximize conversions instead where it is available for your campaign type, which makes it a common first step.
How long does the learning period last?
Google does not give a fixed number of days. Its help article says duration depends mainly on the number of conversions, the length of conversion cycles and the strategy type, and that it can take up to around 50 conversion events or three conversion cycles to calibrate to a new objective. Older advice quoting seven days is not what the current documentation says.
What happens if I set my target CPA too low?
Google's Target CPA article warns that a target set too low can make the system skip clicks that might have converted, reducing total conversions. Start from your recent average CPA and your own lead economics, and change the target in small steps.
Does Smart Bidding fix poor lead quality?
No. It optimises for whatever you define as a conversion. If form fills include many junk or irrelevant enquiries, it will find more of them. Importing qualified-lead outcomes from your CRM, tightening search terms and improving the landing page address quality; bidding alone does not.
Sources: