Adstock: Theory and Easy Excel Template for Marketers

This Blog Post Covers

Introduction

In this blog post, we dive into the fascinating and useful concept of Adstock, a crucial technique in marketing for the success of your advertising efforts. As you progress through this guide, you’ll gain a better understanding of Adstock’s fundamentals and how it can become a valuable asset in your toolkit as a marketer.

Are you ready? Let’s get started 🙂

What is Adstock in advertising?

Adstock (sometimes referred to as “carryover” or “decay”) is a concept first introduced by Simon Broadbent in 1979. This concept holds deep importance for growth marketers.

Picture this scenario: you launch a paid media campaign on Meta or Google Ads, their immediate results look promising if you look at their tyical KPI such as CTR, Impressions, Traffic and Conversions. However, what if the client asks you to analyse the impact of campaigns in the long term? In the other term, what is the influence on consumers even after the campaign has been paused?

Finding out the lasting impact of the advertising efforts is crucial in measuring effectvely the ROI of campaigns that drive higher sales and better performance.

Indeed, Adstock acknowledges that the effects of advertising linger beyond the moment of actual ad exposure. It’s like a ripple effect, where the impact of an ad accumulates over time, leaving a lasting impression on your audience.

There are a range of underlying elements which cause the adstock effect (Willshire, 2022) :

1. Ad Memorability: the higher the emotional resonance in the ad, the more captivating the format, the increased probability individuals will remember it.
2. Purchase Cycle: adstock effect may also be to do with the purchase cycle of a product. For instance, if you come across an advertisement for a product that you only purchase on a monthly basis, it might take a few weeks before you take any action.
3. Channel Mechanism: there’s the channel mechanism to consider. Email campaigns may require a few days for delivery and opening; PPC have a shorter life cycle.
4. Repeat Purchases: consider new vs. frequent customers. A customer may continue making purchases for several months and theoretically, could still be linked to that initial advertising effort.

It is worth mentioning there will also be factors which effectively reduce the adstock such as competitor advertising pressure, promotions, or low customer experience. Also, as you can see in the chart below, some marketing channels tend to have longer effects than others.

Fig 1, Advertising adstock theory, M. Mikes, 2016

How to measure Adstock?

Now, let’s dive into the practical side of Adstock. The formula for advertising Adstock is the following:  

Let’s break it down:

• Adstock at time t, which represents the cumulative effect of advertising at that specific time (week1, 2, 3, or month1, month 2, month 3, etc.)
• Ad(t): Advertising exposure at time x, which is the impact of the advertising campaign in the current period.
• λ: Adstock rate: A constant factor representing how quickly the impact of advertising diminishes over time. It is critical to forecast this value for a reliable outcome.
• AdStock(t-1): Advertising spend and impact in the previous period

When we apply AdStock in a Marketing Mix Modeling, we can represent the value in a time series as a percentage ranging from 0% to 95% which denoted how large the second period (week or month) sales are relative compared to the first period of sales. A 100% Adstock would be unfeasible because it would imply that the positive advertising effect persists indefinitely in the time.

Suppose the advertising leads to 1000 additional sales in the first week, 600 in the second week, 360 in the third week, and so forth. In this scenario, the adstock rate would be 60%, as each subsequent week contributes 60% of the sales from the preceding week. (Willshire, 2022)

As mentioned the most complex step in our model is to find out what the Adstock rate (λ) is? Some analyst may adjust this value by using machine learning models and vendors, other using historical data or bechmark values for verticals; more analysts just using their own expertise, trial and errors. Also is possible to follow documentation in Robyn that helps you to setup specific parameters to improve the model using R.

A Practical Calculation in Excel

Armed with Excel, you can calculate Adstock and gain insights that go beyond immediate campaign performance. Feel free to download the file and edit my Adstock Google Sheet template.

As shown above, in the columns C, D, E I included the actual media spend by weeks. In the columns F, G, H I transformed the costs in Adstock which shows the long lasting effect of media advertisting.

As can be seen in the following chart more clearly for Facebook Ads: by setting a λ of 50%, the positive effect of advertising would lasts for about 12 weeks, while the campaign was running for only 5 weeks.

Conclusion

In conclusion, Adstock is a pratical econometric model that in a privacy-friendly world, has regained its popularity due to the deprecation of thirdy parties cookies and issue with mobile iOS attribution. Adstock is often applied within the Marketing Mix Modeling (MMM) framework to enhance the understanding of the long-term impact of advertising on various business metrics. This enables marketers to tailor strategies based on a more holistic view of consumer behavior and a human centered approach.

Did you find this blog post useful? Do you have any questions? I am happy to schedule a 15 mins of virtual coffee with you 🙂 ☕

 

---

Bibliography

Danaher, P. (2017). Advertising effectiveness and media exposure. In Wierenga, B. & van der Lans, R. (2017) Handbook of marketing decision models. International Series in Operations Research & Management Science 254. NY: Springer.

Firstdigital, (2020, September, 21). Using the Adstock Approach for Modeling Advertising Effectiveness.

Govindan, G., Baig, M. R., & Shrimali, V. R. (2021). Data Science for Marketing Analytics: A Practical Guide to Forming a Killer Marketing Strategy Through Data Analysis with Python. Packt

Mikes, M. (2016, August 27). Advertising adstock theory.

Van Heerde, Harald (2018), Block 2: Measuring Advertising Effectiveness [PowerPoint slides]. Massey University, Return on Marketing Investment. Stream.massey.ac.nz

Willshire, A. (2022). Adstocks: Accounting for the Lagged Effect of Advertising. Recast