Ad-hoc analysis with Google Analytics

Ad-hoc analysis with Google Analytics - justification of traffic increase

Web analysts are often faced with the challenge of providing the board of directors, superiors, customers or employees from online marketing with short-term, numerical answers to facts or anomalies. This blog post shows how a customer’s traffic – without any new campaigns – changed abruptly with the help of the Google Analytics tracking tool.

Step 1: Traffic analysis

To check the initial situation, it was examined whether the traffic had actually increased. To do this, the traffic history is viewed in Google Analytics under the ‘Reports’ tab.

Graph sessions and bounce rate

Figure 1: Partial traffic history in Google Analytics

It can be seen that the traffic in the current calendar week is significantly higher than in the previous calendar weeks and thus significantly above the average of the selected period. In the same step – to get a feeling for the quality of the traffic – the bounce rate was visualised next to the number of sessions.

Graph sessions and bounce rate

Figure 1: Partial traffic history in Google Analytics

Since the bounce rate in this case is slightly below the average value, it indicates that the newly acquired traffic can be of good quality.

In order to further evaluate the quality of the traffic, this step of the analysis also compared whether the number of page impressions per session had increased (as previous analyses had shown that for these customers, the probability of a transaction increases significantly above a certain number of page impressions). Using this defined engagement KPI, it was therefore possible to show that the quality of this traffic had increased in addition to the traffic.

Step 2: Where does the traffic come from?

Traffic has increased, but which source do visitors come from? Using the ready-made report under ‘Acquisition’, then ‘All accesses’ and finally ‘Channels’, it is possible to quickly check which channels are responsible for the increase. For this purpose, it makes sense to select the period before the change with the period after the change via the period selection function of Google Analytics.

Ausschnitt Google Analytics Tabelle

Figure 3: Traffic segmented by entry channel

The channels with the highest traffic in this case are ‘Paid Search’ (SEA), i.e. Google Ads, ‘Organic Search’ (SEO) and ‘Direct’ (direct entry). The strongest changes occur in the selected time periods for the two channels SEA and especially direct entry. The number of sessions for the SEO channel remained relatively constant and can therefore be disregarded. The direct dntry channel in particular has increased significantly by approximately 126%.

Since the direct entries have increased significantly, this is an indication that certain online or offline branding campaigns have worked and the visitors have directly typed in the URL address of the customer in the browser.

In this case, the Google Analytics report under ‘Acquisition’, then ‘Adwords’ and finally under ‘Campaigns’ showed that the increase in traffic for the SEA channel was mainly due to the ‘Brand’ campaign with brand keywords only.

However, this report selection is only possible if the Google Analytics account is linked to the Google AdWords account.

The results from the campaign report were also indicative that the branding campaigns had an impact on product or brand awareness via various other channels and thus visitos typed in the company name as a keyword or a specific keyword in combination with a brand keyword in Google’s search bar.

Ultimately, these indications or hypotheses were verified via a customer journey analysis. In this context, the customer journey analysis was used to allocate the number of transactions to different channels or specific campaigns via a calculated redistribution key. However, this will not be presented in more detail in this blog post.

Note:
  1. For time comparisons, it makes sense to compare numbers not only with one past time span, but with several. If possible, the difference should be tested for randomness using so-called significance tests (a t-test can be useful).
    Further partial analyses not shown here:

Has the percentage of new visitors, i.e. visitors who are visiting the website for the first time, increased (in this case, yes!)?

Graph Sitzungen und organische Suche

Figure 4: New vs. returning visitors

The increase in ‘new visitors’ for the ‘Paid Search’ segment was 75 percent. This was again an indication that the branding campaigns were responsible for the traffic increase. Thus, primarily targeted search engine users called up the customer’s site who had been shown specific terms (incl. images) via offline and other display campaigns.

  • Building on (I), it was further shown that mainly the (new) SEA visitors had entered the keywords displayed offline and via display in combination with the company name as keywords (brand combination) in the search engine.
  • Conversion analyses
  • Path analyses

IT-WINGS NEWSLETTER

Stay up to date!

More blog posts

Analyse funnel in etracker

Analyse funnel in etracker A funnel is a path that users of the website and/or an app follow to reach certain goals (purchase, booking, order, etc.). Funnel analysis helps you

Read more »
two women sitting in front of a laptop

Online Marketing Internship

Internship experience in online marketing Internship experience in the field of online marketing Hello! My name is Alisa and I study advertising and market communication. During my fourth semester, I

Read more »