Let’s imagine the dream of any marketer. You came up with a cool advertising campaign, formulated a strategy, and made creatives and initial customization… And then you click on “Do well” button and the automated system starts. The system collects data, identifies the best placements, frequency, and creatives, and optimizes settings in ad offices. And you just have to observe how the statistics improve and set the parameters that still need to be increased.
It seems like an unattainable ideal, but newage. is already creating such a system, and we plan to start it in 6-9 months. In this article, we will tell you about the first stage of automation.
Now we are finishing the most important stage of creating the system – automation of data collection and visualization. We created a dashboard (access on demand) for analysis, based on 6 years of experience, more than 1000 conducted national campaigns, and more than 10 billion analyzed advertising impressions.
Our reporting automation has become particularly relevant in 2020. When everyone was trying to save as much as possible, we focused on optimizing the resource task – collecting data that made it possible to evaluate the effectiveness of the contact and the advertising campaign itself as a whole.
Dashboards are our competitive edge. In 2019, one report took four hours (this was very fast, given the amount of data that we collected in it), but now it is several minutes and contains even more data. As a result, the customer has all the information about what the budget is spent on and what results are given by certain campaigns.
Our dashboard reflects the agency’s processes based on the Comprehensive Analysis methodology. It consists of four elements:
- quality control of placement and effective contact;
- assessment of media metrics;
- evaluation of the reaction to advertising;
- agile marketing changes depending on the result.
This is how we can assess the real effectiveness of the advertising campaign, but as the method develops, the work processes become more complex, and there is more reporting. After automating these processes, we have reduced the time for preparing reports by more than 10 times.
Step 0. Collecting data about an advertising campaign
Collecting campaign data is a tedious task. A specialist needs to go to all advertising offices, upload information, standardize it, filter it, and only then proceed to analysis. The more tools are used in the campaign, the more time is spent on the task. When launching the smallest display campaign, the report consists of the following data:
- 3-4 ad accounts;
- 2 analytics — Google Analytics and Campaign Manager;
- 4-8 ad formats;
- 15-20 campaigns;
- 20-50 audience segments.
Our reporting idea is to combine all metrics, metrics, and channels into one report using one strategy. We wrote the software and the process has become ten times easier. All relevant information is automatically transferred to the cloud database. We implemented it as follows:
- The main database is stored on BigQuery.
- Load the plan and the structure of the campaign from Google Sheets into it.
- From advertising cabinets that have an API, information is sent to the database via this interface;
- Those sites that do not have an API send reports every day. We upload them to Google Sheets and from there to the database.
- In BigQuery, we combine data under the plan structure and generate reports.
- All collected information is immediately visualized in the form of a dashboard. So both us and the client immediately understand the statuses of the campaign.
Every day, we automatically upload about 100 reports from 12 different sources for 20 advertising campaigns. That’s roughly 1 TB of information. It is difficult, time-consuming, and expensive to handle it manually. Our system already simplifies the task by more than 10 times. Traffic managers see statistics online, in the context of each segment and site. See how bids and budgets are met on Facebook and Google Ads, what conversions and video searches.
If earlier statistics compilation and its analysis took a day of work of a specialist, now the report is compiled every day and automatically. As a result, we control the quality every day, and the quality check of the information received is carried out in 15 minutes instead of 4 hours.
Step 1. Data quality control
After collecting information from advertising offices, we control the quality of statistics and contact. At this stage, it is important to pay attention to the following indicators.
For a complete check, you need to analyze a large amount of information. So we also integrated this process into the automatic dashboard.
Step 2. Assessment of media indicators
Automation has significantly reduced the cost of analyzing media indicators. Below is an example of aggregated advertising campaign metrics from YouTube, Facebook, Display & Video 360, and Trade Desk. We immediately see the aggregated statistics of the overall reach, frequency, and conversions from all sources.
Moreover, the coverage is not just summarized from all advertising offices, it is data taking into account the errors identified in the previous step.
Step 3. Evaluating ad reaction
The most important component of Comprehensive Analysis is the evaluation of ad response. And here it is important to take into account not only interactions with ads (post-click) but also delayed user actions (post-view and cross-device). Below is an example of a report and metrics, the dynamics of which we see every day.
Based on this data, specialists can determine the effective segments and frequency, the most successful creative for a specific audience, and get other insights to improve clients’ advertising campaigns.
It is in this part of the dashboard that unique algorithms of our methodology are used, which allow you to quickly make a decision on campaign optimization.
Step 4. Improvements and the future
In 2020, we reduced the time for preparing reports by more than 10 times, while increasing the frequency of reporting. This allowed us to more quickly assess the effectiveness of advertising campaigns and respond to changes. We plan to develop reporting automation from an automated dashboard into a separate product.
The next step is to make the Comprehensive Analysis processes automatic. The system, according to the given algorithms, will analyze on which site, for which audience, which display advertising gives the greatest impact after contact. The analyst will simply give a command which indicators should be optimized to achieve the desired results, based on his experience and the task of the advertising campaign.
You can request access to the Dashboard example above to recreate a similar system in your company. To do this, send the email you want to access to our mail: firstname.lastname@example.org.