Deeper Dive into Marketing on Amazon

August 18, 2020

Overview

Marketing on Amazon is often described as throwing darts into a dark room. Marketers push buttons and pull levers in a variety of ways, but often the only indicator of effectiveness is Ordered Units, which is provided daily.

While increasing unit sales is certainly one of the main goals of marketing, it would be helpful to have a more informed understanding of the data that drives those sales. Shining a light into that black box would allow marketers to better understand where to focus efforts to achieve a higher return on investment.

This article looks at a specific layer of what is inside the box – Glance Views (referred to as Page Views here) and how that additional data can be actionable.

Page Views and Conversion

For books, unlike other product categories, Amazon does not provide data on page views directly. So Iobyte developed a method to determine the exact number of page views for every product in a publisher’s catalog based on other data Amazon does make available. Let’s look at a simple example, using our fictional title “Dan’s Guide to Effective Amazon Marketing Analysis”.

Suppose that on 8/1, the book sold 10 units.

Using our tools, we know that on 8/1 the title page on Amazon had 100 page views. Knowing the page views, we can now determine the conversion rate on 8/1 was 10% using the simple formula:

Sales ÷ Page Views = Conversion Rate

Img

Measuring True Marketing Impact

On the following day, suppose we put out some ads or perhaps got a mention in an article, and on 8/2 unit sales rose to 15 units. So that is a 50% increase in sales, which sounds nice. But what more could we discern?
Let's look at some scenarios:

Img

In this scenario, the marketing activity was ‘successful’ in that it sent 5x the number of users to the page than the previous day. But a 400% rise in traffic only generated a 50% rise in sales, which dropped the conversion rate from 10% to 3%. The additional traffic generated is lower quality and less likely to convert than the natural traffic the page gets. Therefore, your marketing efforts are generating traffic, but not quality leads.

What you would hope to see is that conversion rate is similar at the higher level of traffic. That would indicate a book where your marketing investment is scaling well, and you might look to dedicate more investment to that same channel. If 10% conversion is maintained, then 400 additional page views would generate not 5, but 40 additional sales.

The extreme version of this is where you create some activity and there is no change in sales. With insight into page views, you can determine whether you failed to drive traffic to the page, or you did drive more traffic and it failed to convert.

Measuring Metadata Change Impact

Another situation page view detail helps with is metadata changes, such as a new cover image, or changes to the book description, price or BISAC codes. Again, you currently only see impact on sales. But a new answerable question arises with these changes – did the change attract more visitors to the page (e.g. Amazon’s search algorithms) who converted similarly, or did the change increase the conversion rate for a similar number of visitors. For that we can compare two models:

Increased Traffic Scenario

Img

In this scenario, the traffic to the page improved with the conversion rate holding steady. This implied a very scalable marketing opportunity to send even more traffic to the title’s page until the conversion rate starts to drop.

Increased Conversion Scenario

Img

In this scenario, the traffic to the page did not change but the conversion rate improved. The net result on sales was the same, but the path was very different. Conversion rate is harder to improve and does not scale well (it gets progressively harder to increase it as it gets higher).

So now we have an additional layer of measurable data that helps us better understand HOW we got to the new level of sales. This new layer helps differentiate between changes you want to invest more into and ones you may not.

Being able to differentiate between these outcomes is highly actionable in terms of what you may want to do in subsequent cases.

Squaring the Rectangle

One final note is to introduce a simple concept to maximize sales using these (or any) two dimensions called ‘Squaring the Rectangle’

Consider a very wide and short rectangle. Lots of page views, but extremely low conversion. Adding more page views – making the rectangle even wider - does not have a big impact on sales because the new page views will presumably convert at the same low rate as the original set. But if you can raise the conversion rate by investing smartly in marketing, that improvement should apply to most, if not all of the current traffic. If it is easier to raise conversion from 4 to 5% than to raise page views from 400 to 500, then that is where you want to invest your resources.

Conversely, a tall but narrow rectangle implies high conversion with low traffic. In that case, you want to drive more traffic to pages which will hopefully retain the high conversion. This should be easier than further raising an already high conversion rate.

In both cases, you have access new data and insights on how to achieve higher sales.

Conclusion

We are demonstrating that adding an additional layer of information to your sales funnel can absolutely help you better understand how you achieve a change in sales volume.

Having the additional data to gain a measure of the quality of traffic sent to your product pages and being able to compare incremental conversion rates as a result of your email campaigns, digital advertising and other efforts is vital in assessing the effectiveness of your marketing and how to optimize it.

This results in the ability to form a predictive model that will make future investment more efficient and return on investment (ROI) higher.

We would like to help you become more profitable and have a clearer view of the selling landscape. Please send questions, comments, or feedback to dan@iobyte.com