A few months into running campaigns, bidding settings usually slide into an awkward state. Manual CPC needs daily attention, and miss a day and the pacing slips. Switching to Smart Bidding brings two scary pieces of advice instead, one saying the algorithm will fail to learn without enough data, the other saying to wait until you have thirty conversions. So the account sits still, budgets either stay frozen or get raised and send CPA upward, and nobody can explain why.

Those hesitations break down into four concrete questions. Is the goal volume or return, is data density enough for a strategy to learn, is conversion value being recorded, and how much volatility can the business absorb. Answer those four and the bid strategy picks itself; skip them and no copied threshold or borrowed benchmark will hold up.

Bidding and budget management has shifted from daily price tweaking to three jobs, choosing the right strategy, managing the learning period, and putting guardrails around budgets. Smart Bidding hands execution-level bidding to the system, while goals, data quality, and scaling pace stay with humans. When conversion tracking and attribution are not solid, automated bidding is a conversation for later; the details live in the conversion tracking and attribution guide, and this piece moves on.

Fix an Outdated Threshold First

Many teams carry one piece of folklore, wait until 15 to 30 conversions before switching to Smart Bidding. That one can be retired. The official guidance organizes rules by purpose, and there are two kinds.

One kind is the requirement to enable a specific strategy. Target ROAS on Search and Shopping campaigns, for example, requires at least 15 conversions in the past 30 days at the conversion tracking level; the bar differs by campaign type.

The other kind is evaluation advice. For Target CPA, the official recommendation is to set the target from the average CPA over the last 30 days, adjusted for conversion delays, and to evaluate over a 30-day window referencing at least 30 conversions.

A third variable decides how long the learning period runs. Google lists three factors, the number of conversions, the duration of conversion cycles, and the bid strategy in use. So a handful of conversions can still get you started, with a different approach, conservative targets, few changes, and a longer evaluation window, rather than idling while waiting to stockpile conversions.

Choose Strategies by Goal and Data Density

The strategy list is short. For volume, use Maximize conversions; for volume at a controlled cost, Target CPA; when conversion value has real spread and return is the goal, Maximize conversion value or Target ROAS; for visibility on Search only, Target impression share. Maximize clicks fits traffic building or low-data transitions.

In common scenarios, lead generation watching a cost ceiling goes to Target CPA; ecommerce with order-value data goes to Target ROAS; a new product launch that just needs presence uses Target impression share or Maximize clicks as a bridge.

Three signals drive the selection. First, whether conversion value data exists and can be trusted, which decides between value-based and count-based strategies. Second, data density, where thin data favors simpler strategies with longer windows over hard-target strategies. Third, business constraints, profit floors, tolerance for volatility, and budget flexibility all feed back into how aggressive the targets should be.

Strategy scope follows account structure, bound to a single campaign or acting as a portfolio strategy that pools data across campaigns. For smaller budgets, pooling data through a portfolio strategy usually beats going it alone and clears the learning period faster. How to build the structure is covered in the account setup and structure guide.

Manage the Learning Period as a Resource

The learning period is when a strategy calibrates to its new goals. Per Google's documentation, its length is driven mainly by three factors, the number of conversions, the duration of conversion cycles, and the bid strategy; once the Learning status disappears, the algorithms keep learning, and manual CPC carries no learning period at all.

Managing it comes down to three rules. First, changes trigger recalibration, so change one thing at a time and avoid touching targets and budgets together. Second, give the window enough room, at least one full conversion cycle, which for long B2B cycles means weeks or months. Third, read progress from strategy status and reports rather than gut feeling. Conversion delay and attribution windows get their full treatment in the conversion tracking guide mentioned earlier.

Budget and Marginal Returns

Start with how budgets work. The average daily budget equals a monthly budget divided by 30.4. Actual daily spend can exceed it, up to about twice on a given day, while the monthly ceiling stays at 30.4 times the average daily budget. Shared budgets let multiple campaigns draw from one pool.

The phenomenon that surprises teams is CPA climbing after a budget increase. That is marginal return behaving normally, since the cheap, strong placements get bought out first and each additional dollar buys less efficient traffic. Judge by whether the incremental cost is acceptable, rather than demanding the average hold still.

Guardrails follow from that logic. Raise budgets in steps rather than doubling outright, leaving an evaluation window after each step; set a profit floor on CPA or ROAS and stop to review when it is crossed; avoid changing budgets and targets at the same time so learning periods do not stack. Shopping and PMax have their own budget dynamics, compared in the PMax versus standard Shopping breakdown.

Platform Updates and Routine Monitoring

Bidding mechanics have kept moving, and two published explainers are worth reading, the Target Bidding update and the end of bid suppression; this piece will not repeat their details.

Routine monitoring covers three items. Strategy status, whether it is learning or limited by budget; the gap between targets and actuals in bid strategy reports; and the periodic search terms and negative keyword review, covered in the keyword-to-page workflow. When the status reads limited by budget, first check whether the increment is worth buying; if it is, raise budgets in steps, and if it is not, put the money where it earns more rather than forcing things by cutting targets. Review weekly, turn anomalies into tasks, and route changes through the same process. The full strategy picture sits in the Google Ads optimization overview.

Common Pitfalls

First, treating a fixed conversion count as a hard gate. Official rules include enabling requirements for specific strategies and evaluation advice, not a universal number; thin data calls for conservative settings and longer windows instead of waiting.

Second, tweaking targets constantly. Every change restarts calibration, and frequent edits keep a strategy permanently unsettled; targets should stay stable across a business cycle.

Third, scaling in one leap. Marginal cost rising is the norm, and doubling budgets in a single move usually backfires; step up with guardrails and watch the increment.

Fourth, setting Smart Bidding and walking away. Learning status, budget limits, and target gaps all need eyes; the system takes over execution, but judgment stays with people.