Quality Score in Google Ads: What It Consists Of and How to Improve It
Quality Score in Google Ads rates a keyword from 1 to 10 using three components. We explain what it means, what it affects, and which actions really improve it.
Manual and automated bidding strategies in Google Ads solve different tasks. We explain which to choose at the start of a campaign, what the learning period is, and when to move to target CPA or ROAS.
At the start of a new Google Ads campaign you usually choose a strategy that collects data: manual cost per click or maximize clicks if there are no conversions yet, and maximize conversions if tracking is already working. Once enough conversions have accumulated, you move to target CPA or target ROAS.
A bidding strategy sets the rule by which Google decides how much to pay for a click or an impression in each auction. It defines what exactly you ask the system to achieve: more visitors, more leads, or more revenue. The metrics you analyze depend on it as well.
Strategies are divided into manual and automated. With manual ones you set the bid yourself, and with automated ones the system picks it for each query individually, relying on signals: device, time, location, search query, and history.
| Strategy | What it aims for | When it fits |
|---|---|---|
| Manual CPC | Holds the cost per click you set | Small budget, full control, first tests |
| Maximize clicks | The most visits within the budget | No conversions, you need first traffic |
| Maximize conversions | The most conversions within the budget | Conversions are set up, there is little data so far |
| Target CPA | Conversions at a set average cost | There are stable statistics and a clear acceptable cost per lead |
| Maximize conversion value | The largest total value of orders | A store with different order sizes and value passing |
| Target ROAS | A set return on ad spend | There is order history with value and stable sales |
If tracking is not set up or the campaign is only launching, an automated conversion strategy has nothing to learn from. Here manual control or maximize clicks with a maximum bid limit is appropriate. In parallel, set up tracking of leads and purchases so you can move to automation later.
When the first conversions appear, you can move to maximize conversions. The system learns from real results and redistributes bids between queries. For stores with different basket values, maximize conversion value is the next step.
Then you add target CPA or target ROAS. The goal is set based on the real history of the campaign. A threshold that is too strict limits impressions, and the campaign simply stops spending the budget.
After a campaign launch or a strategy change, the system collects data and selects bids. During this time the results are unstable: the cost per click and conversions can change sharply. This is normal if the campaign already has a minimally sufficient amount of data.
Do not change the strategy because of a short-term dip. Seasonality, competitor promotions, or weekends cause fluctuations that are not necessarily related to the settings.
Automated strategies work only on quality data. Check that conversions are set up correctly, that identical events are not counted twice, and that an action that brings no money is not counted as a conversion. You also need negative keywords and a clear campaign structure, otherwise the system learns from junk queries.
A store launches Google advertising with a small budget. At first conversions are tracked, but there are few of them, so the campaign runs on maximize clicks with a bid limit, and the main work is cleaning queries with negative keywords. When leads and purchases become regular, the campaign is switched to maximize conversions, and they watch how the cost per order changes.
Next, the store passes the real amount of each order to Google. Now you can try maximize conversion value: the system favors queries that bring larger orders. If sales are stable and it is known what return on ad spend is acceptable, target ROAS is added, first with a soft goal that is gradually tightened. At every step one thing is changed and the date is recorded.
An automated strategy needs room to learn. A daily budget that is too small relative to the cost per conversion gives too little data for conclusions, and the system cannot find patterns in time. If the cost per lead in your niche is high and the budget is modest, it is better to start with narrower campaigns: one topic, one region, clear keywords. Then even a small number of conversions points in one direction.
Also separate products or services with very different margins into different campaigns. For cheap and expensive products the reasonable goal differs, and a shared target CPA will be too low for some and too high for others. Separate campaigns let you set separate goals and compare results fairly.
After moving to a new strategy, compare the period with the previous one of the same length, and better still take seasonality into account. Look at several metrics at once: the number of conversions, the cost per conversion, impression share, spend, and revenue if it is passed. If there are more conversions but the cost has risen, the question is whether this remains profitable for your business.
The last point is often underestimated. An automated strategy optimizes for what you specified to it as a conversion. If both valuable and random leads fall into that category, the system cannot tell them apart without a hint. That is why it is useful to pass the value or to import from the CRM only those events that actually led to a sale.
No. They need data and clean conversions. On a small budget or in a new niche, manual control often gives more control.
There is no clear universal threshold, it depends on the campaign. Look for a stable flow of conversions over several weeks and for hints in the account itself.
Usually up to a few weeks, depending on the volume of traffic and conversions. During this time avoid big changes.
Yes, in different campaigns. For example, proven queries can run on target CPA, while new topics are tested with maximize conversions.
Check whether the target CPA or ROAS is too strict, and whether geotargeting and keywords are too limited. Often it is enough to soften the goal a little.
The bidding strategy should match the stage of the campaign: data collection, learning on conversions, then management by target cost or return. Without correct conversions no automation will help. If you want a team to handle the setup and review of results, LightSEO can help. See how Google advertising works, and what role a website audit plays before a campaign launch.