Integrated promotion of your project

Why conduct a test run ppc
advertising campaigns

Couldn't it be simpler, and what do statistics have to do with it?

The test campaign is the most important. Without it you can not answer the question - is it even possible to use contextual advertising to promote your site and receive orders from customers

Is there no way to know that
in advance?
Before preparing a promotion proposal, a PPC specialist conducts an analysis of search demand for your topic. It is possible to predict the level of traffic by key phrases in the search for both Yandex and Google with the help of special tools - planners
прогноз бюджета google ads
прогноз бюджета яндекс директ



посчитать расходы на контекстную рекламу
Still, it is better to calculate at least order of magnitude of expenditures and a preliminary
estimate the effect than to START wasting BUDGET IN VAIN
These calculations don't tell you exactly how many impressions, clicks, and requests from converting customers you'll get at that price. Forecasting tools work with a large margin of error, and the data in the media plan are not specific to your site. Yandex has rules for the planner: take into account the average performance for all selected phrases in the selected region, the site will evaluate the data from current advertisers, and it will be the average values for the past 28 days, not for the next period.

The main flaws in the Google planner algorithm: rounding average values; the more frequently a key phrase is used in a search, the greater the error in the calculation of traffic (high-frequency phrases may have an error in frequency of display of tens of thousands); Google planner hides some low-frequency keywords. Conversion rates in the preliminary calculations also do not reflect the data specifically for your site. These are average figures.
As I accumulate data, I calculate average conversion values separately by topic, and for those industries where such data is insufficient, I use the average values published in statistical studies. Another important thing that makes forecast calculations inaccurate is the quality of campaign setup. On how well an ad will be set up, greatly depends on its effect. Quality advertising from different advertisers can vary greatly, it depends on what specialist carries out the setting up and maintenance. Depends on the conversion and how well made and how interesting the proposal on the site for your customers.

All this leads to the fact that the discrepancy in the data evaluation and data from the real launch will be inevitable
If there is no other way, then
How long does the test take and how much will it cost?
The test period lasts from 30 days or more, depending on the niche. If you spend less time, the results will be inaccurate and you will not be able to judge anything from them. An exception to this rule and that doubtful - test immediately on a large budget (from 50 000 ₽ and above). But we don't know what the result will be, so it's not a very smart decision. Plus there will still be the question of exactly how demand changes over the month and beyond. Plus

It is necessary to confirm the forecast data on the real launch of campaigns

there was It is proven that it takes 2 weeks or more to run their algorithm. The main task of the test - to get a statistically valid sample of data, on the basis of which you can judge whether it is possible to scale up advertising: to increase the budget and expand the semantic core. So it is worth having patience, so as not to waste money, and the budget for the first run is the one that was calculated for you by a professional.

Run the test to answer the question: how many requests from the site will be for every 100 conversions from advertising

In order to begin somehow to plan the result, you first have to understand: how many requests will make visitors to the site with every 100 hits on the site. This data will help to make slightly more accurate calculations for the future. If we proceed from this goal in the test, on average, the minimum traffic that is needed - from 100 clicks. For each case, the size of the optimal sample can be calculated separately, and it will help to get more accurate results. To do this, there is a special calculator

The calculator will help you calculate the amount of traffic you need for text-based campaigns.

1. Enter the size of the population to be studied (general population)
2. Specify the confidence level you want (usually 95%)
3. Enter the confidence interval (value in %, tenths, dot).* Then click the "Calculate" button
* - The confidence interval is the maximum permissible error of future results, e.g. ±3%.
The main thing you get after the test - an understanding of whether promotion with the
Contextual advertising, not "by eye", but confirmed by real data
Made on