Search marketing has always been defined by the access to a wealth of information which enables data-driven decision making. However, the latest trends within automation are changing how marketers work with data. In this final post in our Search 2020 series (read Part 1, Part 2 & Part 3), we will focus on how data-driven decision making has changed within Search marketing and discuss how scientific methods, trend tracking, and breaking down data silos in a privacy-complaint way will enable marketers succeed in the coming years.

Data-driven decision making has always been the bedrock of how Search campaigns are run, but what happens when marketers can no longer perform valid experimentation on granular levels? Marketers used to be able to steadily enhance performance by fine-tuning creatives through iterative experimentation and optimizing CPC by gradually exploring the bid landscape. Those days are over. 

Additionally, as few queries happen without the user having some predisposition towards certain brands, marketers must acknowledge that they will work with an incomplete data set. Other digital channels and traditional marketing (OOH and TV) together with the limitation/deprecation of cookie-based tracking, leaves marketers with a major blindspot in terms of the efficiency of their marketing activities. 

Marketers need to be very aware of the limitations of the data they work with, and especially be concerned with over indexing the effectiveness of their marketing activities and the impact of their own interventions. However, there is still plenty of value in the insights collected from Search marketing activity. In this post, we will dive into the three factors marketers must account for before insights can be translated into actions.

Experiments: Go Big or Go Home

What makes Search marketing special is the immediate feedback we get from users. They either click or do not click on an ad. It is that simple. The format of the channel requires the user to pass judgement immediately and this creates an abundance of data and trackability unchallenged by upper-funnel counterparts like video and programmatic marketing. Working with such a data-rich channel opens up many opportunities, such as being scientific in experimentation. 

One of the complicating factors nowadays is that rigid A/B testing is not feasible, since Google will often pre-select the audience used for certain activities. They will, for example, not show the same ad copy to different audiences, and sometimes certain ad copies aren’t shown at all. This selection bias greatly hampers marketers ability to conduct valid experiments, even with the existing features available in Google Ads.

Marketers thus need to be aware of the limitations to how they can use their data. For example, micro-decision making is often best left to the algorithm, which can account for non-linear correlations and compute infinitely more signals at all times. Therefore, Search marketers must let go of granular testing and instead focus their data-savviness on bigger experiments. Such experiments could be switching-off bidding for brand terms or tailoring ad copies to the geographical location of the user. Search marketers need to have the guts to conduct bigger and riskier experiments if they want to ensure that account performance continuously improves.

First rule: Experiments need to be less granular and should be used to challenge even fundamental ideas about how Search marketing campaigns are run.

Transfer Insights Across the Company

Carrying insights from Search marketing activity to the rest of the business might seem like a low value task. However, Search marketing is in many ways the most direct source of information about customer sentiment. Typing in a query on Google is nowadays the first step users take after realizing that they have a need. This moment is what Google refers to as the Zero Moment of Truth, or ZMOT. In 2014, Jim Lecinski described ZMOT as:

… the precise moment when they have a need, intent or question they (read: consumers) want answered online.

Search queries themselves hold intrinsic value since they carry the information about what potential customers are searching for. For example, using intelligent labeling to track the surge of key themes or product searches will enable spotting of important demand trends which can be actioned by all parts of a firm, from the Financial Planning manager to the Supply Chain planner. 

Another potential case is using Search marketing for the testing of new USPs before companies launch major marketing campaigns. Developing creative material for richer media such as video and display comes with high cost and risk, so CMOs will naturally welcome an opportunity to see how users react to changes in the company’s communication before taking the leap. Bulk changes to text ad copies are a comparatively very low effort undertaking.

The value of the insights that can be derived from Search marketing has broad utility, so marketers need to proactively ensure that this information is tracked and presented to the right people to enable timely and effective actions to be taken. 

Second rule: Insights about the ZMOT should be transferred to other parts of the organization to ensure that the value of insights is spread.

Breaking Down Data Silos the Right Way

Yet another problem facing marketers nowadays is the continuous fragmentation of the data they need  to make the best investment decisions. The emergence of increasingly strict data protection legislation and a general rising societal awareness of privacy. 

Marketers need to understand what their privacy policy allows them to track, while also being aware of any inherent blindspots embedded in the data. They also need to make sure that they are prepared for a future where first and third party cookies are much more fragile than they are used to.

This leaves marketers with a completely different data foundation; one which 1) will contain far fewer user identifiers and 2) less data points in general. It is important that marketers do not see this as a blocker for utilization of data in their day to day work, but rather a change in the nature of the data. This is especially important for cross-channel decision making as the channels are likely to have some disparity between them in terms of the data richness. It is critical that marketers can identify those differences and make coherent decisions on how data from different channels should be assessed. 

The next step is the establishment of a “Single Source of Truth”, i.e. a complete, quality-assured, collection of data, which can become the bedrock of impactful data projects such as CLV-bidding and customer segmentation. 

But wait? Didn’t you just say that data will be more fragmented?

Yes, and no. Although marketing data overall will be more fragmented, an updated, future-proof data collection strategy will naturally quality assure that the data marketers have access to. The data foundation will improve thanks to increased transparency and accountability, which is why investments in proper data integration solutions will only increase in value going forward. 

There is no doubt that marketers need to embrace privacy. But with a better foundation, marketers should be able to properly analyze the behaviour of the opted-in users and embrace the power of aggregate over individual data. This will help them break down the data silos and make better decisions about marketing investments across the organization.

Third rule: Marketers should embrace privacy and focus on using quality assured data for defragmented data about the user journey.


With the rise of algorithmic data usage marketers need to be even more aware of the limitations, but also the opportunities, that exist in the data they have access to. If marketers remember the three rules for translating data insights into actions, they will be equipped to continuously improve the performance of the Search marketing campaigns and also to create value more broadly across the business.

Andreas Toth Arentoft Senior Data Specialist