Maximize Conversions and the missing cap on click price

You switch a campaign to maximize conversions. Inquiries grow. Then one morning you open the report and see a single click that cost as much as ten ordinary ones. You go looking for the setting that caps the price of a click, and there is no such setting. This is not a bug in the interface. On this strategy the ad system takes the price of every click for itself and spends it according to its own forecast of value.

Below: what you hand over with such a strategy, why cost per lead cannot be set directly even with manual bids, and what to manage instead. The original piece was about Google Ads. For a company selling into Russia the platform is Yandex Direct, so I keep both in view.

Why there is no field for a click price cap

Automated strategies that optimise for conversion volume set the price of each click on their own. The logic is simple: the system estimates what a given impression is worth, taking its value forecast into account, and bids whatever it considers right. A field where the advertiser could limit the maximum cost per click does not exist in such a strategy. Google Ads works this way with Maximize Conversions, and the automated conversion strategies in Yandex Direct follow the same principle.

The direct consequence is that you have no built-in tool to cut off one-off expensive clicks. In practice this shows up as a handful of clicks that cost many times more than the account average, on a tiny number of impressions. On a large, even budget a couple of such outliers dissolve in the total and change nothing. On a small budget, or in a narrow niche, a single anomaly can skew the averages for a whole period. You end up reading numbers that describe a random spike rather than your campaign.

This matters more for foreign companies entering Russia than for local advertisers, because new entrants usually start exactly where outliers hurt most: a test budget, one or two narrow product lines, no account history. The first month’s report is the one that decides whether the market gets a second month, and it is the report most exposed to a single expensive click.

Manual bids will not give you cost per lead either

The obvious fix seems to be leaving the automated strategy for manual bids and holding the click price yourself. There is no direct lever there either. It is simply hidden deeper.

Even when a group of similar keywords shares one bid, each keyword behaves in its own way. It has its own threshold at which it enters the auction at all, its own range of actual click prices and, most importantly, its own range of prices where most of its real conversions sit. Conversions for a particular phrase do not always sit at the cheap end of that range. Sometimes the opposite is true.

When you change a shared bid at the level of a group, a campaign or an account, every keyword under that bid enters or leaves the auction at once, and the one keyword you meant to adjust is only one of them. Standard reports do not show how many phrases started or stopped receiving impressions after the bid change. As a result the shift in cost per conversion cannot be cleanly tied to the change in click price, because it is mixed with a change in the set of keywords taking part in the auction.

Third-party bid managers that tune bids towards a target return face the same wall. If the click price at the level of an individual query jumps around almost at random, no algorithm will hold cost per lead through the bid.

In Yandex Direct the effect tends to be louder than in a Western Google account. Russian queries are heavily inflected, one intent is spread across many word forms and phrasings, and the keyword sets that grow around a single product are wide. A single shared bid therefore moves a larger and more mixed crowd of phrases in and out of the auction. If you want to see how that plays out in bids and forecasts, I cover it in bids in the Yandex Direct auction.

The only model in which cost per lead is controlled directly and predictably is paying per conversion. On search, though, traffic volume under this model is usually very small. That narrows it to tasks where volume is not critical, or it has to be mixed with other sources. Yandex Direct offers this option inside its automated strategies, and I describe its trade-offs separately in pay per conversion in Yandex Direct.

What to manage instead

If neither the automated strategy nor manual bids give you a direct lever, the economics are held indirectly, through several supports at once:

• A budget limit caps how much the system may spend in a period and stops one expensive day from dragging the whole month with it. Google Ads budgets are set per day. Yandex Direct automated strategies are typically budgeted per week, so the protection works over a longer window and a bad day can be absorbed within the same week. Plan your monitoring rhythm around the period your platform actually uses.
• A target cost per conversion gives the strategy a reference point to build bids around instead of chasing volume at any price.
• Campaign segmentation isolates risk. An anomaly in one narrow campaign does not poison the statistics of the entire account.
• Regular monitoring looks for outliers by hand before they pile up. A glance at the statistics every few days catches an expensive click before it becomes a trend.

This is not the same direct control as a max CPC field that does not exist. It is management of the frame inside which the system is free to make decisions. The shift in thinking is this: you do not dictate the price of each click, you set conditions under which a rare expensive click cannot wreck the economics of the period. A single outlier does not always spoil the overall picture. On a small budget or in a narrow niche, though, it can distort the averages noticeably, and the frame is what insures you against that.

For a company managing a Russian account from abroad, one practical addition: agree in advance who looks at the account and how often. Automated strategies punish long gaps in attention, and a campaign that is checked only when the monthly invoice arrives has no frame at all, whatever the settings say.

Read the signals that click attribution does not show

There is one more support that tends to be forgotten in discussions of automated strategies. When exact attribution of every click is impossible, decisions are made on combined signals rather than on one cell of a report. Specialists learned this discipline long ago from offline advertising.

TV, radio and digital screens do not deliver clicks directly, but they create delayed demand. It surfaces online as growth in brand and navigational queries and as more direct and organic visits. Click attribution is powerless here, so people rely on indirect evidence. They watch branded search after a flight starts, prepare the sales team for a wave of inquiries in advance, and ask customers where they heard about the company. A small budget drowns in the noise of organic demand, while noticeable reach produces a visible wave.

Exactly the same logic applies inside search automation. If the system hides the price of each click from you and gives no clean link between bid and cost per lead, read the wider picture: movement in branded demand, the share of direct visits, and feedback from sales on the quality of inquiries. Look at the sum of signals around the metric instead of the metric alone.

In Russia the tools for this are local. Branded demand is watched in Yandex Wordstat rather than Google Trends, and it is worth knowing what Wordstat shows and what it hides before drawing conclusions from it. Visits and goals are read in Yandex Metrica. Sales feedback often comes from messenger conversations, which many foreign teams do not count as inquiries at all. If those conversations are missing from your picture, the quality signal you are reading is incomplete.

It also helps to synchronise efforts in time. When offline and digital channels launch out of step, their contributions overlap chaotically and it becomes even harder to see what worked. When display, outdoor and search advertising run in a coordinated way, the combined effect on awareness grows, and a spike in branded demand is easier to tie to a specific period. Seasonality and parallel campaigns can still mask the contribution of a single channel, so keep the accompanying factors in view instead of crediting all growth to the most recent launch.

What a foreign company should do differently

Three points are specific to selling into Russia from another country.

First, do not port targets from a Google account. Google stopped serving ads in Russia in 2022, and whatever history you have describes an auction you no longer take part in. Set the budget limit and target cost from your current unit economics, and let the Yandex account build its own history before you tighten anything.

Second, split campaigns earlier than you would at home. With a thin Russian site and a new account, segmentation is the cheapest insurance you have against one outlier deciding the fate of the test.

Third, make sure the signals you plan to read are actually collected: Metrica goals, call tracking, and a way to count messenger inquiries. Without them, the wider picture this article recommends does not exist.

What to do with your campaign this week

• Check which strategy the campaign runs on and whether you have a field for a click price cap at all. If you do not, the risk of outliers has already been accepted by default.
• Set the frame: a budget limit and a target cost per conversion, so the system does not chase volume without a reference point.
• Move directions with different risk into separate campaigns, so that one anomaly does not distort the overall statistics.
• Build the habit of looking at expensive clicks and abnormal days every few days instead of waiting for the monthly report.
• Do not reduce the assessment to one figure from the ad account. Keep branded demand, direct visits and sales feedback next to it.

Maximize Conversions honestly does what it promises: it brings conversions. In exchange it takes the click price out of your hands along with part of the transparency. You can accept that deal, but accept it with open eyes, before the expensive click has already been charged.

If you cannot see where the budget leaks on an automated strategy, or why the averages keep jumping, book a free review and I will show where the risk sits and what to fix first. For how automated strategies fit into a paid traffic system for Russia, see the Yandex Ads page.

Frequently asked questions

Sources

• Google Ads Help, about Maximize conversions bidding: support.google.com/google-ads/answer/7381968
• Yandex Direct help, bidding strategies: yandex.com/support/direct/en/strategy
• Russian version of this article on belokrylovo.ru: Максимум конверсий и лимит цены клика в Google Ads