Post completely revised and updated on June 24, 2021
Which sales would have happened without any marketing efforts?
How can fluctuations in sales figures be explained? To what extent does competition influence our sales figures?
Many advertisers ask themselves these and similar questions. Because the fact is: no matter how good and effective your advertising is, it is not solely responsible for all sales.
Consumers are subject to a multitude of different influences, which can affect their buying decision. They often act out of habit. If they have had a good experience with a brand or a product before, they are very likely to revert to the same brand for their next purchase.
Sometimes we shop simply because we feel like it or because the new iPhone that we've been waiting for so long is finally on the market. The influences and impressions that affect us every day are sheer limitless. Depending on the study, the number of ad impressions we receive alone ranges from 5,000 to 10,000 per day.
To fully understand the incremental value of marketing, marketers must be able to measure and evaluate the relevant influences that lead to a conversion. Conversions that would have taken place even without marketing campaigns form the baseline. The demand that goes beyond the effects of the baseline is marketing-induced. In this post, we have summarized the most important aspects you need to know about baseline modeling.
Baseline Modeling in a Nutshell
The baseline is a model that captures external effects that cannot be attributed to marketing. In the standard model, this includes the following components (cf. Fig. 1):
Structural demand: This forms the basis of the baseline. It contains the fundamental level of demand. A large component of structural demand is brand-induced. A good way to increase structural demand is therefore to invest in the company's brand and image.
Seasonal demand: In our model, we define seasonal demand as any recurring fluctuations in the intensity of demand in the market. These systemic changes can be weather-related but also caused by vacations, public holidays, and other cycles within a year (e. g. release of a new iPhone, Black Friday, etc.). They have a massive impact on customers' preferences and consumption decisions and lead to a lump formation in the demand volume.
Trends and disruption effects: Conversions originate not only in demand but also in supply. Thus, trends are also considered when it comes to conversions. Here, it is assumed that trends are subject to change and that their intensity can change over time. For example, the introduction of a new product variant can lead to exponential growth in conversions through word-of-mouth effects.
Fig. 1: The Components of Exactag's Standard Baseline Model
Conversely, it means that extreme declines in demand can occur in times of exogenous shocks (such as the coronavirus crisis), (cf. Fig. 2).
Fig.2: Disruption Effect Demand Collapse in the Exactag Baseline Model
Exactag calculates the incremental contribution of marketing using fully automated econometric models. In the process, the focus is on the causal influence of marketing activities on conversions. If this causality can be statistically proven for a certain proportion, this proportion of conversions is allocated to the corresponding marketing activities. The proportion of conversions that is influenced by external effects and cannot be causally attributed to any marketing measures is attributed to the baseline accordingly.
Besides the components of the Standard Baseline (structural demand, seasonality, trends & disruption effects), other important aspects can be included in the modeling. Exactag's Custom Baseline allows for detailed analysis of elements such as brand awareness, competitive spend, pricing, etc.
Why is the Baseline so important?
For advertisers who want to measure and optimize their marketing campaigns in the long term, the baseline offers two fundamental advantages:
The influence of marketing can be calculated precisely: Advertisers who know how their baseline is evolving can accurately determine the impact of their marketing campaigns. For example, if sales have increased by 25% in the last quarter, advertisers can use baseline modeling to narrow down the success more precisely. This may show that while the baseline increased, 15% of the increase in sales still came from a new marketing campaign that was launched that quarter.
Explain fluctuations in performance: As the baseline algorithm progresses, advertisers become increasingly able to categorize and evaluate performance fluctuations. For example, if sales drop despite a constant marketing budget, these fluctuations can be explained by external factors such as seasonality, trend or disruption effects, or increased competitive pressure.
Brand image as an important factor in the Baseline
The brand image reflects whether a brand is the First Choice with consumers or at least part of the Relevant Set when it comes to brand choice. Is your brand the one consumers are instinctively drawn to when looking for a particular product? Merely having heard of a brand versus buying from a brand makes a serious difference.
Advertisers put a lot of time and marketing effort into creating their brand. Investing in brand image, or increasing brand equity, is, therefore, one of the most time-consuming and cost-intensive tasks in marketing. This effort is justified because companies with an excellent brand image can often justify higher prices than their competitors.
However, being the first choice in consumer brand selection depends on other factors besides the brand image. Prices, offers at the moment of demand when consumers want to buy, also play a major role. Consumers may ultimately be persuaded to buy another brand if it offers a supposedly better deal at the point of demand. In general, "First Choice" can therefore be influenced in the short term by advertising activities.
Integrating Baseline Modeling into your marketing attribution
In multi-channel marketing, the development of a baseline is essential for the accurate analysis of marketing activity effectiveness. A baseline can be used to discover relevant insights for each channel, which in turn can be used to optimize campaigns.
Data-driven attribution accesses data from the standard baseline model (or custom baseline model) to take into account the relevant external factors. The parameters of the baseline model are automatically calibrated on a weekly basis. Key questions of Exactag's Baseline Model are:
- How many conversions were realized in the previous week?
- What incremental contribution did marketing make to the company's success in the period under consideration?
Attribution analyses thus become much more accurate, and the actual value contribution of a channel or marketing campaign can be determined with complete data. Advertisers can thus use their marketing budgets in a more optimized way to ensure marketing and business success in the long term.