Step 3.10: Analysing Traffic Data
Intermediate SEO Concepts | Analysing SEO Data
Website traffic data comes from analytics programs like Google Analytics.
In this step we introduce some concepts surrounding the analysis of the traffic data we introduced previously.
Analysis Over Time
Knowing what date range to look at is often a first step in looking at data… Sometimes it won’t be an option, because the site has only been tracking data for X amount of time. Often, an SEO will be asked to look at a specific date range as therein exists a problem that needs unpicking.
No matter what or why you are looking at data, picking your date ranges will be pivotal to the insights you can produce. There are several considerations when looking at data over time that you must consider or understand:
- Seasonality - plays a huge part in traffic volumes and conversion rates
- Market Trends - often dictate how you can expect to perform
- Tracking - if this was not setup properly, you may not have accurate data
The first two can be mitigated by looking at long time frames (12 month or higher) where trends and seasonality should average out.
Compare Like for Like
When looking at traffic data or performance data of any kind, you need to compare like for like. This includes ‘errors’ such as comparing all traffic in Feb to Organic traffic in March, these are not the same and don’t make good comparisons.
More often what we mean by this is comparing valid date ranges with one another. Comparing December-16 to January-17 is often a pointless comparison due to aforementioned seasonal factors becoming more dominant drivers of performance.
Typically, we make year on year comparisons, to see how well Feb-17 is doing compared to Feb-16 & feb-15. Clearly this relies on at least 13 months’ worth of data to be possible. Although management and clients all want to see MoM (Month on Month) growth, more often than not YoY (Year on Year) growth is a better performance indicator.
To tell a story with data, you often need to derive valid correlations between multiple datasets. The most common would be ranking data and traffic data; changes in rankings should correlate to changes in organic traffic.
This is just part of the story, if rankings go down, so does traffic… But why? This is a very common problem and question for SEO’s. to answer it, we need to bring in other data such as Google Algorithm updates, competitor data, trend data, website data, etc.
We need to look for correlations between these many data sets, for example; if Google released an update that had the potential to affect a site, and within 2 weeks of this the site saw a drop in organic traffic and rankings… The story starts to write itself!