If you were to google “conversion rate defined,” you can expect to get about 2.2 million results, most of which will state something like “the percentage of users/visitors who take a desired action.” But does that make it the right definition?
Think about the “desired action” part of the definition, as it seems to leave a great deal of room for interpretation. For instance, a retail clothing company desires a very different kind of action from its customers than, say, the DMV. The relative nature of conversion rate inherently poses a problem for a one-size-fits-all definition, but all too often we see marketers thinking of conversion as a singular and absolute calculation that can be applied to any and all channels within its business.
There are as many different ways to calculate conversion as there are products, all of which tell a different story and have a valid reason for being considered in analysis. Needless to say, it’s easy to go down a rabbit hole when it comes to measuring conversion, so for the sake of being concise, we’re going to focus on measuring conversion for retailers in the digital space. We’ll take a deeper dive into the framework of conversion and why it’s important to understand how conversion rate is calculated. Later, we’ll discuss different conversion rates and touch on which might be most appropriate for measuring your desired output.
Let’s start by first staking out a few easily found conversion pitfalls:
Industry benchmarks don’t always apply. Avinash Kaushik put it best by saying “all data in aggregate is crap.” This same sentiment can be applied to industry benchmarks. Who is defining your industry? Are there factors unique to your company that significantly affect your ability to convert visitors? Wouldn’t you agree that all companies in your ‘industry’ have unique factors of their own? There are so many variables influencing the outcomes of a company’s digital performance that it’s hard to justify pooling them all together to generate a universal benchmark by which to assess your performance. My suggestion? Benchmark against your own historical performance first, and then supplement that analysis with select industry benchmarks for context.
Data sources are not created equal. Conversion rates are dependent upon how your analytics platform calculates metrics like visits, orders, visitors, pageviews etc. You will find some discrepancies when comparing the number of visitors to your site reported by Google Analytics and IBM’s Coremetrics because each interface calculates a visitor differently. As such, a conversion metric like buyer/visitor rate will be impacted by which source of data you use (this will be discussed in greater detail in part 2 of this post). Investigate how your web analytics platform calculates key metrics. If you need to align it with previous data, most platforms allow you to adjust the way a metric is calculated by changing settings or making simple code modifications. Ultimately, you’ll want to make sure that the source behind your conversion rates is consistent, and that over time, the calculation of the metrics used in that conversion rate remains constant.
Where to attribute the credit? There are any number of different attribution models and all of them have their merits, but they all assign the value of a conversion differently. For example, the “last interaction” model attributes 100% of the conversion value to the last channel the customer interacted with before buying or converting. Conversely, the “first interaction” model attributes 100% of the conversion value to the channel with which the customer first interacted (Google Analytics does this). Clearly, conversion by channel is greatly influenced by the attribution model used.
As illustrated by the graphic above, you can even customize how your analytics tool assigns conversion value. By doing this, you can specifically design a model that fits your assumptions in evaluating conversion path data. Additionally, many of these tools also include a feature allowing you to compare how different attribution models affect the valuation of your conversion paths. By comparing multiple models simultaneously, you can gain further insight into how different models are evaluating conversion by channel.
This is an example from Google Analytics’ Model Comparison Tool. Here, we can see how the Last Interaction, First Interaction, and Linear Attribution models all stack up against each other in an easy view. Dig deep into how and when to use each model, because they all have their uses. Google has some great literature on attribution models and there appropriate usage here.
In summary, as with all key performance indicators, the most important point is to know where your data is coming from and how it’s being calculated. Developing a keen understanding of the dynamics of conversion can make you a better marketer and help you avoid common mistakes in analyzing your data. Later this week, we’ll drill into different conversion rates by channel and how to use them to tell the most accurate story in your data.