Marketing Science, what is it and what could it be? (Part 3)
Welcome to the third part of this series. Before continuing let’s do a quick recap of what we already covered in Part 1:
- Role of Marketing Science Teams based on the maturity of the organisation.
- Different Marketing Strategies that can be supported: acquisition and retention.
- The possibility of Marketing Science to contribute across the entire customer lifecycle.
- Industry-wide focus on advertising spend measurement and optimisation.
And in Part 2:
- Media Mix Model, Attribution Model, and Lift Tests for measuring and attributing marketing spend. Strengths and limitations of each tool.
- Introduction of marketing terms: lag effect, attribution window, saturation.
- Emphasis on Marketing Team’s ownership of budget, channels, and strategy. Marketing Science’s role in developing measurement tools.
In this post, I aim to share my experience with channel-specific measurement and performance analysis, striving to keep it as channel-agnostic as possible. The emphasis will be on online channels, as digital marketing partnerships typically provide good amount of data.
Disclaimer: While this post could be considered more aligned with marketing analytics than marketing science, one cannot sprint without learning to walk. This serves as an essential step before progressing to the development of machine learning models for marketing.
The key to comprehending the performance of your marketing investment lies in having a clear view of the Marketing Conversion Funnel. Personally, I identify four stages in the funnel that deserve analysis: Impressions, Clicks, the Check out Process, and Conversions.
To draw a parallel between the aforementioned stages and a traditional marketing conversion funnel, it would look something like this:
- Impressions serve as the Exposure of our product to audiences, aiming to generate Awareness.
- Clicks represent the stage of Discovery and Interest by our prospective customers.
- The Check out stage is the initial clear sign of Consideration and Decision by our prospective customers.
- Finally, Conversions encapsulate the Action we are hoping our customers to take.
However, this funnel is not static, as it mainly depends on your marketing spend. Therefore, I prefer to include marketing spend at the top of the chart and transform the funnel into a time series view:
Ideally, there should be a positive correlation across all the charts: the more you invest, the more impressions you get. This increased presence in the media channel should translate into more clicks and traffic to your site. The new traffic generated should lead to more check-out processes, ultimately resulting in a higher number of conversions. However, this is not typically what occurs.
I like to refer to this as delta analysis: take two points in time, measure the difference (delta) across the various stages of the funnel, and analyse whether your budget decision was accurate.
For example, if we analyse the delta of day 7 compared to day 6, we invested only 4.74% more budget, which generated 21.98% more impressions, 31.71% more clicks, 35.02% more checkouts, and finally, 20.76% more conversions. Definitely a good decision! Encouraged by this outcome, you decide to keep pushing on day 8, but then you encounter a surprise: your additional 6.56% spend generated only 1.42% more impressions, and clicks decreased by 10.72%. Fortunately, the checkouts increased by 20.94%, and the conversions went up by 4.34%.
In the example we randomly generated the data points and we were not able to observe the effect of seasonality, something we must know from our business. But this is a quick and easy way to understand if our budget decision was a good decision and identify where we should be focusing our follow up analysis.
Stage 1 — Impressions
Continuing with the example, if we take a look at the chart of spend and impressions from day 5 to day 8 we can observe that impressions kind of flatten. This could be a signal of reaching the saturation point in terms of audience.
Be aware Marketing behave according to the law of diminishing returns, a principle which states that benefits gained from something will represent a proportionally smaller gain as more money is invested in it.
In order to identify if the saturation of the audience is reached, some partners provide some metrics called Impressions Share, Top Impressions Share and other derivates, which helps us understand how much of the audience we are still be able to show our ads.
Stage 2 — Clicks
Another scenario that might arise is when the delta of impressions surpasses the delta of clicks for the same time period. This situation can be elucidated through the performance marketing bidding process. In the context of a bid in digital marketing, it signifies the sum of money that a company is prepared to offer an advertiser to display their ad. However, contrary to common belief, money alone doesn’t dictate success in the marketing realm.
In the performance marketing bidding process, three primary components come into play: the amount designated to pay for displaying the ad, the ad’s relevance to the user’s search, and the overall quality of the product being advertised.
Even if there’s a substantial increase in the amount allocated for each bid, resulting in a higher marketing spend, without a concurrent enhancement in ad relevance or quality, more ads may be displayed, but clicks won’t follow suit. Users won’t be interested unless there’s an improvement in the relevance and quality of the advertisements.
If this situation is happening in our business then we should pay attention to our ads creative, product feed and also review the search terms for which we are showing ads, in order to exclude those which are not relevant or not performing according to out business.
Stage 3 — Check out
The next delta to analyse is the one between clicks and check out, and I really like to make clicks and check out instead of clicks and conversions. Just wait for it.
What we are looking for is that the increase of click translates into the increase of check out processes. If we only get the first part of the equation, then we were able to awake the Discovery sense of the prospect customer and the Interest to check our products, but we failed on make the customer arrive to a Decision. Checking the bounce rate could give us a hint if we have to review our landing page.
Stage 4— Conversions
The ultimate goal of marketing is, of course, generating a purchase, but in my opinion, what happens between the start of the check-out process and the conversion is really a product responsibility. If we are marketing a great product and our platform has a smooth check-out process, then we should not observe a decay between check-out and conversions.
However, this is not always the case. A misaligned delta of the check-out process and conversions is the worst-case scenario for a marketer, as it is a part of the funnel that does not depend exclusively on Marketing. During the check-out process, our customer reviews our full value proposition, and if any surprises come up (usually hidden costs), most probably we are going to lose them. My suggestion in this case is to collaborate with the Product team. Review the check-out process stages together and understand the reasons our prospects are leaving the check-out process. Once we identify the possible reasons, try to incorporate them in order to create a transparent marketing communication.
You are not alone
The funnel and delta analysis can be conducted at different granularities of your marketing structure: channel, group of campaigns, campaign, and sometimes even at the ad level.
As mentioned earlier, this is not a static analysis; it evolves and not exclusively with time. There is another component we haven’t mentioned yet, but it is always there: competitors. You might think this is obvious, but in digital and performance marketing, competitors play a crucial role. Remember that each ad impression we are exposed to is the result of a bidding process. Impressions Share can help you understand a bit about the competition in your audience. Some marketing partner also make available a Clicks Share metric. But before using it, read the details of how the metric is being calculated by the partner.
Are we really managing marketing spend?
So far, we have been discussing the funnel, and we assumed that we were influencing it with the management of our marketing budget. However, with the main marketing partners, this is not exactly how it works, as they have sophisticated mechanisms in place to manage your budget. These are known as bidding strategies.
A bid strategy determines the way you bid in paid advertising platforms for your ads to show up in available spaces
Let’s be clear about the trade-off: current bidding strategies make the management of our marketing budget easier, but they also make it less transparent. This presents an opportunity for Marketing Science to assist in managing the full funnel: translating your bidding strategy into estimated budget spend and later into the rest of the stages is crucial to effectively manage your marketing campaigns.
A comment about the translation of the bidding strategy parameter into marketing budget and later the full funnel: your marketing partner usually sells you a click with a probability of conversion. When can the conversion occur? That depends on the attribution window you have set in the partner platform. So, keep this in mind, as dismissing the attribution window could give you an incomplete view of your marketing performance data.
Opportunity
As you can see, digital marketing is a vast data world that someone has to take care of. Marketing teams are able to do it; they master it! But the scale in the number of campaigns, the multitude of metrics to review, and day-to-day operations require marketing teams to get some extra tooling.
Marketing Science is the partner to develop what in the industry is called Ad Optimisation, which includes bid optimisation, keyword optimisation, creative optimisation, marketing landing page quality score optimisation, audience targeting, among others.
I personally believe that understanding the full process of marketing is a must before starting to develop data solutions. I always encourage my teams to be aware of the industry they are working in. In my humble opinion, that is the key to success and business impact.
I hope you enjoyed reading the third article of this series. In the following publications, I would like to switch the focus a bit from media channels to our beloved customers, those who commonly go unnoticed when marketing teams are fully focused on channel management. Please leave comments about your experience with data and marketing to enrich our collective knowledge!
