That “first-party data is important” is a universal axiom in modern marketing. Commonly understood to be consented personally identifiable information (PII) obtained by marketers from customers through compelling value exchange, it is thought of as the fuel that powers addressability to customers across the purchase lifecycle and the most valuable currency accepted by the flagship privacy-preserving audience solutions rolled out by major ad platforms as a response to increasing privacy regulation and the decreasing viability of third-party identifiers.
But that is just one definition of first-party data. In the context of a marketer’s larger business operations, different parts of the organization may view “first-party data” differently. They may view it as the data assets about their individual customers, acquired from their respective business processes, and which live in their own respective siloed systems: ad campaigns, content marketing, CRM, billing, transaction records, point of sale, customer service, site and app analytics, loyalty programs, inventory management, social media, email management, and so on. And modern digital marketing is not just targeting ads to well-constructed audience lists; it is the deployment, at scale, of customized experiences to customers at every digital touchpoint to foster a relationship with the brand. Ironically, every one of these customized experiences can in turn be fueled by precisely the siloed data assets collected by the rest of the organization.
How does the digital marketer realize the full potential of these data assets and fulfill the promise of modern digital marketing? The answer ends with key partnerships with platforms and service partners who facilitate activation of use cases across each touchpoint, whether it is display ads, or video, or social media, or email, or SMS, or site personalization. But it begins with a strategy prioritizing the centralized management and orchestration of every type of first-party data, a strategy that breaks down silos and unlocks digital tech maturity for the marketer and the rest of the organization.
Centralized Data Strategy: Organizational and Technological Enablers
Centralizing data assets from the myriad of a business’ systems to democratize data is a noble goal. But the process to achieving that goal is a multifaceted one, requiring enablement both organizationally and technologically. Organizationally, businesses must align to ensure that there is executive sponsorship, an agreed upon philosophy of when to partner and when to build, and a strategy for skill development. Stakeholders from marketing, information technology, CRM, and other teams need to be aligned on the strategy, and best practice is for these stakeholders to form a central working group with established points of contact to bring forth the collective data vision for the benefit of the entire organization.
Technologically, the leading cloud platforms have enabled this goal by offering centralized data warehouse solutions that manage first-party data. All leading cloud platforms offer 1.) scalability and efficiency through elasticity and interoperability, 2.) security through identity and access management (IAM), encryption protocols, firewall rules, secrets management, denial of service and web attack protection, the ability to partition data confines vertically and horizontally, region-specific data location options, and well documented shared responsibility models, and 3.) bundled analytics capabilities to enable business stakeholders to easily query the centralized data and deploy it for marketing and non-marketing use cases. Kinesso has found these capabilities invaluable, especially in the context of establishing trust with clients whose first-party data we work with to deploy advanced marketing capabilities. Protecting the data of our clients, and of their underlying consumers, is a foundational principle at Kinesso and our cloud platform partners enable us to do this at scale globally.
Data Standardization and Unification
Because different types of first-party data come from a myriad of systems and arrive in diverse formats, orchestrating the ingestion and standardization of that data through thoughtful Extraction, Transformation, and Loading (ETL) processes is critical in order to make it useful for the marketer. The leading cloud platforms offer ETL tools that simplify data ingestion from external data sources, which enable marketers to approach their data warehouse stack in an open-source manner and handle much of this critical work in-house.
But whether managed in-house or with the assistance of a partner, implementing ETL best practices to ensure a performant data warehouse is just the beginning. For brands with many disparate sources of first-party data, some partners can offer customizable ETL solutions that streamline and automate this process for a brand’s unique systems infrastructure, even for non-advertising data sources like point of sale, loyalty programs, and CRM. A smaller subset of partners like Kinesso can even enrich the datasets to fill gaps, further unify datasets with proprietary identity resolution services and first-party identity graphs to create a 360-degree view of customers, enable activation of this data across a wide range of advertising platforms, and implement sophisticated measurement of activation activities.
Using Centralized Data to Power Advanced Marketing Use Cases
Once data is centralized, standardized, and unified in a data warehouse it is democratized for use by marketing teams for a wide range of applications. Kinesso has successfully deployed these use cases for a wide range of clients, demonstrating success in advancing their digital maturity and driving business outcomes.
Customer 360: First-party data takes a variety of forms and comes from a variety of systems. Scalably managing the ingestion of this data from a multitude of systems each with nuanced export processes and transforming it from a multitude of formats into a usable form is a foundational component of creating a centralized and unified data repository. But the disparate customer touchpoints and accompanying systems where data originates inherently creates gaps in data capture. Records from a customer call center may be comprised predominantly of phone numbers. Records from point of sale may be comprised predominantly of credit card numbers. Records from email subscribers may be comprised of individuals’ work or personal email addresses depending on the nature of the underlying product being offered. Records from site visits before authentication may be comprised of pseudonymous first-party cookies. Records from mobile applications may be comprised predominantly of mobile identifiers.
To realize the value of each of these data points, these disparate records need to be unified. Identity resolution solutions play a key role in this process, filling the gaps from each disparate touchpoint to create a 360-degree view of each individual customer. Kinesso’s identity resolution solutions can solve for this and even append additional demographic, behavioral, and other attributes to further enrich these profiles and unlock further insights and activation opportunities.
Predictive Modeling: The continued decline of third-party identifiers means the decline in viability of audience solutions that rely on those identifiers. In their place, marketers have an opportunity to leverage their centralized and unified first-party data in conjunction with advances in predictive modeling to fill the gap.
Predictive modeling is a mathematical technique that analyzes a set of input data and applies statistical methods and machine learning to forecast future events. High-fidelity first-party data, with metadata from the joining of the company’s other datasets and a 360-degree view, can be a powerful input for predictive models. For marketing use cases, predictive modeling can be used to either foster relationships with customers about whom a brand has detailed information, or to find other potential customers and begin fostering new relationships.
For instance, with Propensity to Convert, marketers can take individuals identified as leads from a key conversion event before the purchase and predict how likely they are to ultimately make a purchase, based on previous onsite interactions and the various metadata from their 360-degree profiles. Splitting these leads into cohorts of likelihood tiers, marketers can then concentrate media spend on the highest likelihood cohorts and drive higher sales. Similarly, for brands whose revenue relies on repeat customers, Predictive Lifetime Valueallows marketers to concentrate spend on newly acquired customers that are likeliest to be repeat customers, tailor customized promotions to them over digital channels, or even focus their initial prospecting efforts on individuals who are most likely to be long-term customers. Churn Prediction can be used for brands that rely on recurring purchases (such as subscription-based companies) to predict which existing customers are most likely to lapse by analyzing previously observed behavior patterns of customers who have already lapsed. The marketer can then reach those pre-lapse customers with advertising, app notifications, email promotions, custom content, and product updates to drive reengagement with the brand. For brands that have implemented CRM, it is straightforward to segment users based on the pre-purchase interactions tracked in the CRM. But Smart Consumer Segmentation can play a key role in segmenting users on demographic or behavioral attributes. While a travel brand may know that a user subscribed to an email newsletter, they may not know which of those customers are interested in urban destinations, or which of those customers may be interested in cultural excursions once they get to their destination. By matching first-party data with a robust consumer database, employing predictive score-based indexing techniques, and layering in CRM-tracked attributes, brands can create granular, highly qualified cohorts from their first-party data that can be activated directly or used as seed audiences in lookalike models.
But to realize the full potential of Predictive Modeling, it must be linked directly to activation. In the absence of third-party identifiers, this requires a flexible and customized model development approach that considers the ultimate intended activation platforms, and which will generate outputs that are accepted by those platforms. Increasingly, this takes the form of segmented lists of hashed PII with metadata, scores attached to individual platform-specific pseudonymous identifiers, or calculated values assigned to pre-purchase conversion actions which can be used to inform custom bidding algorithms. Kinesso in particular employs a data unification strategy and data science approach that enables predictive models which accept more types of first-party data inputs, are able to generate more forms of model outputs that can be activated in more platforms, and are constructed with a built-in feedback loop that re-informs the models with the campaign results that those activations generated.
Customer Journey and Personalization: With the power of a centralized data repository, marketers can use Customer Journey Analysis to map their customers’ interactions with the brand across digital touchpoints and over time. Customer Journey Analysis helps brands understand which journey sequences lead to the strongest customer relationships, diagnose the interactions that leave users frustrated, identify gaps between what the brand thinks it is delivering vs. what the end customer actually experiences, and maximize the revenue potential of each customer purchase. Informed by Customer Journey Analysis, a brand can then take the next step of orchestrating and activating customer journeys using platforms that specialize in each type of digital touchpoint from advertising to email to SMS to site personalization.
An online retailer may employ Customer Journey Analysis to determine that most high-LTV customers purchased incremental products based on recommendations shown to them throughout the onsite experience of evaluating their primary product search, and low-LTV customers are referred directly from search engines and convert relatively quickly post-referral. With this knowledge, the brand could implement additional pre-checkout recommendations based on a scoring of visitors (informed by their 360-degree view) to maximize the transaction value and incentivize customers to return to the site. Further, the site experience of future return visitors could be personalized to show additional products that are commonly paired with their initial purchases.
To realize the full value of Customer Journey and Personalization, Kinesso has developed a robust analytical toolset and utilized frameworks grounded in hypothesis testing. We have also undertaken integrations with leading activation partners that are able to orchestrate journeys from our analytical findings on behalf of our clients.
Measurement and Insights through Dashboarding and Advanced Reporting: With a centralized data warehouse unifying data from various platforms and modern data visualization tools, marketers can generate dashboards to spot trends and manage performance of their advertising campaigns, updated in as real-time as data ingestion schedules allow. Drawing from datasets from multiple platforms, these dashboards can report on custom calculations and metrics that are not available natively within the individual platforms. Dashboards can also supply business intelligence for other parts of the organization that the marketing team may report to, and can be provisioned to various stakeholders with appropriate read and write permissioning. Importantly, with centralized data from offline sales, online conversion, and digital customer interactions (both paid and owned), and with the assistance of conversion measurement enhancement tools like Google’s Offline Conversions and Enhanced Conversions, marketers can build sophisticated reporting capabilities that measure true lift in business outcomes driven by their campaign investments.
To accommodate the varied business requirements of brands in varied verticals with varied reporting stacks, Kinesso has made a concerted effort to develop experience in data management across a wide range of data sources, compatibility with a wide range of visualization tools, and expertise in deploying custom attribution methodologies.
The Path to Maturity: CDP vs. Component-Based Approach
Digitally mature marketers have approached the acquisition of these capabilities in different ways. Many leading Customer Data Platforms, which are themselves cloud-based, offer these capabilities out of the box; for some large, global brands with disparate legacy systems an investment in a CDP is a logical choice.
But for other brands, such as smaller digital natives without the resources to onboard a large enterprise-level CDP, a component-based approach that exploits the inherent elasticity of cloud platforms and is flexible enough to grow at each brand’s unique trajectory toward digital maturity may be the right choice. Kinesso has successfully worked with brands to deploy both strategies, acting either as a CDP implementation and service partner for some brands or as a solutions provider delivering individual advanced cloud-based marketing capabilities for other brands. Our unique organizational integration with agency media planning and buying teams, as well as the extensive platform expertise across our network, has enabled us to cohesively deploy these services from end-to-end: implementation, to analysis, to activation, to measurement.
While most of the leading cloud platforms offer tools to build these capabilities, Kinesso has found success deploying these solutions in Google Cloud Platform (GCP), particularly for brands with investment in Google Marketing Platform (GMP). GCP is deeply integrated with GMP, offering a suite of tools with native connections that streamline data ingestion, data export, and visualization of media activity from GMP products.
BigQuery simplifies data ingestion with Data Transfer Service, which provides native data pipelines from major marketing platforms as well as other non-GCP cloud storage and data warehouse solutions. Through the Google Cloud Marketplace, there are also many connectors built by third parties available to support data ingestion from platforms that do not have native pipelines to BigQuery. Because many brands may want to work with a partner to deploy first-party data solutions, BigQuery also enables easy cross-organization data transfers with Authorized Views, which enable brands to selectively share datasets with outside parties on a need-to-know basis through robust permissioning and data restriction options.
GMP enables activation of predictive model outputs with a robust catalog of APIs that can be easily deployed in GCP. Customer Match can be used to activate model outputs consisting of hashed PII, and approved partners like Kinesso can use the Customer Match API to automate this process. Outside of PII, many brands are realizing the value of predictive models for site interactions leveraging durable Google Analytics 4 (GA4) data in preparation for Universal Analytics’ sunsetting in 2024. Depending on the brand’s GA4 configuration, these models can be designed to generate predictive scores attached to individual client IDs or user IDs, which can then be uploaded back to GA4 at scale and in automated fashion using the Google Analytics Management API’s Data Import feature. For activation on non-GMP platforms, BigQuery can also enable the export of model outputs to their supported endpoints through solutions like BigQuery Omni.
Looker is itself a powerful tool within the GCP suite for a number of use cases. Most Looker use cases leverage it as a seamless data visualization tool for datasets housed in BigQuery and other data warehouses, and it excels in these use cases with pre-packaged visualizations, prebuilt integrations, and a robust set of APIs to support new integrations. But it is also a web application development platform, enabling dashboards that also serve as tools natively integrated to GMP where marketers can directly deploy campaign optimizations and create and upload audiences. For instance, Looker recently released an integration with Customer Match that enables first-party lists to be uploaded directly from Looks and Explores to Google Ads.
Conclusion
Brands collect a wealth of first-party data in various forms from its various consumer touchpoints, and this data is a marketer’s most powerful asset in the path to digital maturity. But to realize its full potential, brands must pursue a thoughtful strategy of centralizing and unifying the data, enriching it to fill any gaps inherent in each touchpoint’s unique collection process, and deploying it for advanced marketing use cases.
Each of these steps can be enabled by leading cloud platforms, and each can be done either in-house or with the assistance of key partners that add unique value with proprietary solutions for each step. Kinesso has assisted brands in advancing through each step of the data-powered digital maturity lifecycle, leveraging our proprietary data and analytics solutions, white glove service offerings, and the powerful synergies between Google Cloud Platform and Google Marketing Platform to deploy advanced marketing capabilities driven by customer data. In evaluating the platforms and partners to work with, brands should consider the key advanced marketing use cases they require and the combination of technologies and services that can flexibly grow with them on the path to digital maturity.
