Exact adjustment of contextual advertising rates depending on the time of display [+ template]

As usual, the bids are adjusted: they figured out that the output of traffic / conversions is less than two times less, and they reduced the rate by half - well, by eye. And then a little impressions, conversions or with a budget, something is not right.

For lovers of accuracy, we offer a ready-made solution: a template for adjusting bids in Yandex.Direct and Google Ads. Download the data on the distribution of traffic / conversions by time from Google Analytics, insert into the template and get adjustments. Everything is simple and the main thing - without the “knee-knee” calculations.

By the way, our traffic specialist really likes the template.

Logic for bid adjustments based on delivery time

Bids are adjusted based on traffic or conversions.

Traffic-based adjustment is suitable in two cases:

• Your task is to save traffic and save on its cost;
• you have very few conversions (for example, a young business or an expensive product), which is why the data for adjustments have a high error.

How is the savings achieved? At peak days / hours, due to the increased activity of competitors, auctions overheat, and to get good positions, you need to set higher bids. The lower the audience’s activity, the less money you need to pay to get a unit of traffic.

In each field of activity, fluctuations in activity are unique. For example, when selling goods for newborns, hot time is at lunch and in the evening (when the children are sleeping and the mothers are “sticking” on smartphones). And for manufacturers of valves, the peak falls on the pre-dinner time on working days.

Conversion-based adjustment allows you to display ads at more profitable positions during the hours and days of the week when users more often buy / order services / register, etc.

For business, this option is a priority adjustment, but also associated with problems:

• Set up goals or ecommerce
• conversions should not be single, otherwise the adjustment will be random.

Next, we show where to get data for bid adjustments and how to get ready-made adjustments using the template.

Where to get traffic / conversion data

Traffic

Google Analytics allows you to identify traffic fluctuations. Open the “Audience” / “Browse”, set the hourly display of data, set the date range and export the data to XLSX.

As for the date range: we set them in such a way as to upload data for the last 4 full weeks. For example, if today is August 12th (Monday), then we are not unloading the current week, but unloading the previous four from Monday to Sunday (from August 15 to August 11 inclusive). This is important for the correct display of data in the template.

Why 4 weeks, if adjustments are made in the context of one week? Everything is simple here: we calculate the arithmetic average of attendance in order to smooth out possible monthly fluctuations in demand.

After export, open the report and go to the "Dataset1" sheet. If everything is done correctly, then the last value in the “Hour Index” column will be 000671.

Data on traffic fluctuations is also in Yandex.Metrica. Go to “Reports” / “Standard reports” / “Attendance”, set the period (defined in the same way as for Google Analytics) and hourly detail. We upload the chart data in XLSX format.

The report should have 672 lines with data (lines 7-678 are filled).

Which data to use - from Yandex.Metrica or Google Analytics - is up to you.

Conversions

If you decide to adjust bids based on conversions, then the data is downloaded from the system in which goals or electronic commerce are set up (depending on which indicator you use as the basis).

For example, if you have goals set in Google Analytics, go to the "Conversions" / "Overview" section, set the details by day, date range (also 4 full weeks) and select the target of interest.

Export data - it will be in exactly the same format as when exporting traffic data.

If, for example, you have configured goals in Yandex.Metrica, open the "Conversions" report and click on the target you are interested in. A report on targeted visits will open. Set detail by day, set the period to 4 full weeks, and upload the data in XLSX format.

Three problems with data collection

Problem 1. Counters are not set or goals are not configured.

Solution: set the counter, set goals and collect data. If the deadlines are on, then at least one full week. And if there is time, collect data for the month.

Problem 2. Severe seasonality.

For products like sunglasses or winter tires, it makes no sense to analyze the previous month if it is off season.

Solution: collect data for the last year (select the peak period and upload statistics). Well, of course, if you just set the counters (see problem 1), you will have to wait until peak periods are determined.

Problem 3. Little traffic / conversions

It’s good if your site is visited and you can collect enough data. And if there is no traffic / conversions? Here the analytics system will not help - it will have nothing to collect.

This problem is the most unpleasant.

There are several solutions:

• Do not set bid adjustments, wait until statistics on transitions from the context accumulate, and only then adjust them.
• Minimally adjust "by eye". For example, it is obvious that if you have a site of an industrial enterprise, then there will be a decline on weekends and at night. For this time, set the adjustment at 70% of the maximum bid. And during working hours, do not adjust the bid.
• Use indirect methods to determine audience activity. For example, find a group on social networks that is similar in theme to your site (if you have your own active group, this is generally good). Analyze the group using the Popsters service - there is data on the distribution of activity by time of day and day of the week. Yes, this is not accurate data, but the overall picture is approximately clear.

Which way to choose, see the situation. But if we talk about accuracy, it’s better to collect data. Prior to this, either do not adjust the rates at all, or adjust them minimally during periods of obvious drawdown.

Bid adjustment using a template

Copy the bid adjustment template for yourself. On the "Initial data" worksheet, in the yellow area, insert the data about the traffic (or conversions) from Yandex.Metrica or Google Analytics. You don’t need to organize the data - just copy them.

Go to the bid adjustment sheet. In the blue area, ready-made adjustments for each hour by week (in% of the maximum bid).

Let's see how it all works.

In the section of the table “Input data”, the arithmetic average of attendance on a specific day and hour (for 4 weeks) is calculated.

The data on attendance are divided into 4 groups using quartile:

• for hours in which traffic exceeded quartile 3 (that is, 25% of the most peak hours), set the maximum rate - 100%;
• for hours in which traffic exceeded quartile 2 (50% of the most visited hours), we specify an adjustment (in our example, 90% of the maximum bid);
• for hours in which traffic exceeded quartile 1 (75% of the most visited hours), we specify an even larger adjustment (in the example - 80%);
• for hours in which traffic was below quartile 1, we set the maximum adjustment (in the example, 70%).

You can change the percentage of adjustments - in the yellow area on the sheet "Bid adjustment". Advertising systems allow you to reduce bids up to 90%, but we do not recommend this, because there will be little traffic. In practice, you should not apply an adjustment below 40-50%.

Important! Quartile-based adjustment is just one option. You can change the template and adjust bids based on the median (there will be two types of bets - maximum and reduced), quantile (there will be maximum and 4 adjusted bets), or any other way of dividing the sample into groups.

Set ad schedule

So, the percentage of adjustments you have. It remains to make them in advertising systems.

Yandex.Direct

Temporary targeting is set at the campaign level. Open the campaign settings and in the "Delivery schedule" section click "Change." In the window that opens, turn on the bid management mode and transfer the adjustments. On the “Result (for Direct / PromoPult)” sheet, the data is adapted for easy transfer to Direct.

Please note that in Direct you can set an adjustment above the maximum bid. Using the template, you can make an upward adjustment without problems (for example, set 120% instead of the maximum bid at 100%).

Google ads

Here, the time targeting is also configured at the campaign level. To do this, open the campaign of interest, go to the “Advertising schedule” section and set the schedule for the display.

The problem is that Google Ads does not have such a convenient visual chart as in Yandex.Direct. Manually setting a complicated schedule is not easy.

If you know how to use scripts, then the Sophie Newton solution will help you. Fill in a special template with adjustments for Google (take data from our template) and connect the script.

If the scripts are not for you, use auxiliary services. For example, when launching a context using the PromoPult platform , time targeting is configured in the visual editor mode for both Yandex.Direct and Google Ads (simultaneously in two systems).

Achtung: time targeting!

Setting up a show schedule is a good thing. If you do it right. Things to keep in mind:

• If you are faced with the task of saving traffic and saving a budget, adjust bids based on traffic;
• if you need sales / registrations and other useful actions, adjust bids based on conversions;
• do not turn off impressions during drawdown periods - better adjust your bids downward;
• if you have a seasonal product, take data on the analysis of audience activity for the peak period last year;
• if there is little traffic data to make a decision on the adjustment, it is better to postpone it until statistics are accumulated;
• if you have few conversions, but a lot of traffic, adjust based on the traffic;
• test different adjustment schemes (for example, one week - 100/90/80/70, the second - 100/80/70/60, the third - 100/90/70/50, etc.) - so you will reach the optimal level budget savings.