The world of marketing has changed in how we talk to consumers, segment, measure and understand performance. With regulations, range of data & technology partners and GAFA (Google, Apple, Facebook, and Amazon) have changed the marketing landscape on its head. Brands need to re-think the role of data & tech in marketing and build an ecosystem that can drive the business forward with a data strategy at the heart of it.
Working on a consultancy project end of 2020 for an e-commerce retailer they wanted to understand what their marketing data & technology ecosystem should look like. The e-commerce retailer is going through digital transformation, looking to make significant structural changes in how they operate, growing their internal team and capabilities.
Developing a data & tech ecosystem sounds exciting but not having the right culture, structure, and team it becomes an expensive project. It ends up becoming a hindrance within the business not being able to get the most out of their tech or better understand the data.
The first phase was to get a better understanding of the business and what their main areas of concern were:
- Tied into the Google and Facebook ecosystem
- Integrations of additional technology into the wider stack causes friction
- Combining different data sets becomes a challenge
- The current set-up and data availability are unable to answer business questions
- The current measurement does not align to the business
- Got a rich data set and insights on customers which are not being utilised
What this translated into, two core areas that had to be the focal point:
- How can I get more control and ownership of my data?
- How can I get more out of my data?
One decision that was made was to move away from the Google 360 stack based around:
- Concerns over privacy how Google use the data
- The inability to work seamlessly with other data & tech partners outside of the Google stack
- Not being able to have ownership of the data
- Insights from Google Analytics pushing to spend more within the Google media network
How an independent ecosystem can answer my challenges
Building an ecosystem should be sustainable and overcome any challenges that can be combatted allowing for key business challenges to be answered:
- Build a data & tech infrastructure that facilities flexibility and growth
- Owning your data, you become less reliant on technology
- 360-degree view of my customers
- Data led segmentation and targeting
- Develop a measurement framework aligned to business goals
- Modelling can be done with the current data set or adding additional data to the mix
Building the marketing data & tech ecosystem
When building a data & technology ecosystem it should be built with the idea it can be plug and play and the impact on the business, data and performance should be minimum. On average a brand has 91 different martech tools but the core ones that every brand needs to have are:
- Data Lake
- Data Warehouse
- Tag Manager
- Single source of truth
- Customer data platform
Having the core tech in place makes the process of integration, available data, data mapping a much easier process. In total there are 15 different data & tech platforms that are needed to build the core marketing ecosystem. The plug and play approach provide great flexibility to work with any partner at any time with the core tech stack ensuring there is no negative impact on the business, data, and performance.
The biggest challenge when deploying new tech, it is not given the sufficient time so most of the time its half-baked and does not deliver the value it was promised. The whole end to end process of building out the ecosystem I plan it would take a period over 3 years.
How can I get control and ownership of my data?
Investing into a Data Lake and Data Warehouse
The quality and trust of data has taken a beating over the last few years. Mainly down to these 3 reasons which helps provide a compelling business case why data control and ownership is a necessity.
Privacy:
- GDPR and CCPA laws stating consent is required before tracking
- Safari, Firefox and Chrome all having their own privacy laws and the death of the cookie
Measurement:
- Google and Facebook changing how they are both validating conversions and changing measurement to suit them
Investing into a data lake and data warehouse provides the ability to integrate different data sources together from marketing data to CRM to make better business decisions.
The four key areas that a data lake and data warehouse can help:
- Prove business value
- Access to raw data (not only marketing data)
- More control over data quality
- Flexibility over measurement and analysis
Within marketing you end up working with multiple data & tech partners and you have access to bottomless pit of data which does not get touched or analysed. Collecting raw data from different data sources available is key to generating deeper insights and better measurement to help prove the value of marketing within the business. It allows the business to move away from softer metrics such as sessions, conversion rate and CPA to Profit, Retention, Churn and Lifetime value. This is not possible by using one source of data, but multiple data sources will have to be stitched together.
This is what potentially a data ecosystem will look like sending different data sets to the data lake and data warehouse
Before creating the data pipes, you would want to have an exploratory data discovery session. With this session, the following can be raised:
- Can the business questions be answered with the all the datasets available?
- How it can be mapped together from identifying primary key?
- What kind of insights are possible?
Looking at how the data cleaning / data governance can be improved to help the process. With storage costs not as dear it gives the option to go with ELT (Extract, Load, Transform) process instead of ETL. The ELT option you can maintain the raw data within the tables to manipulate later.
Another important area to look at is server-side Tag Management which can help solve the privacy issues.
Building the base – Tag Management (TMS)
Tag managers play a significant role in the ecosystem. Having a solid and robust TMS set-up is the start of the journey for brands in building a data & tech ecosystem. From audits that I have done I have been shocked by many of the TMS as it seems that for most sit in the corner. I see this as a good indicator how data is valued in the organisation and I struggle to see i.e. how a CDP will be managed.
Easily forgotten that a TMS will help power a core part of the data that is being sent into the CDP. If the tracking and measurement from the TMS cannot be trusted that immediately devalues the data that is going to be shared with the CDP.
This is what potentially a TMS ecosystem looks like with all the different data & tech partners reliant on the TMS.
Building the base – Single source of truth
When building data & tech ecosystem there needs to be selected platforms that uses a single source of truth. The single source of truth will be the platform that will be sending raw data into the CDP and the data lake. Selecting a platform that can provide raw data is key in building out the ecosystem.
Having a well refined set-up of the single source of truth capturing all the key data points on a granular level that can answer key business questions will help power the CDP and data lake.
How can I get more out of my data?
Customer Data Platform
In the current world it’s becoming a challenge for brands to identify and know who their customers are. This is where CDP’s have become attractive as they can help combat the challenge around 3rd party cookies being no more putting a focus on first party data (customer data).
The CDP is the brains behind the marketing, and it can push data driven decision making when executing campaigns.
The main areas that a CDP can help with:
- It can ingest variety of data sources
- Move away from data silo’s
- Create a single customer view
- Seamless activation across platforms
- Create a persistent customer profile
The core part of the CDP is the data that it can ingest. It becomes extremely easy to ingest all data sources into the CDP and the integration process is seamless. You need to be selective about the data sources you ingest coming from your 1st party data + single source of truth being your primary. Then adding in the likes of mobile app data or paid media platforms data if required and LTV data at certain points.
CDP can help prove value by:
- Delivering marketing campaigns at scale and speed with the right segmentation
- Time efficiency around data integration and process
- Suppression, getting more value out of paid media budgets
- Leveraging customer data to build sustained long-term relationships and drive higher margins
To understand the value the CDP is providing it does require a good understand of the business and marketing operations:
- How are my segments currently performing?
- How are my marketing channels delivering v targets?
- How much time and value is being spent on:
- technical integrations i.e. over 12 months
- on creating campaigns
- creating segmentation lists and working with different platforms
This will help assess pre-CDP v post-CDP the impact on the business.
When working with new shiny tech it’s very common to try and get the most out of it ASAP and prove the value. What happens most of the times that the tech is not fully integrated to get the value out of it. In my planning I have gone for 6 months integration of data sources in a 24-month plan broken down into 4 different sections:
- Integrating data sources to CDP
- Customer data and insights
- Advanced Analytics
- Creativity and Campaign Automation
Proving the value
With a sizeable project like this the cost investment is significant and it will need to prove the financial value it will provide back to the business. Some key steps involved to help with the process of proving the value:
- For each technology selected map out the positioning of each data & tech vendor against value, cost, and importance. It will make it clear to the business the value that each tech provides and where a business case will be needed. For example: a CDP the value has been justified but the cost is significant, and it takes a long time to implement it will need a business case to get the CDP approved.
- With this kind of project part of a wider transformation it would need its own cost benefit analysis which is likely to be looked at over a 5-year period.
- For each technology in the ecosystem getting a good understanding of its costs, how each data & tech provider charge if its recurring or non-recurring. This will help when aligning costs with benefits.
- When creating a cost benefit analysis, the headline metric that will be looked at how long will it take to break-even and make a profit. It will be different for each business and the model they have. Having these insights available will provide the right kind of storytelling that is needed. Got to remember that marketing is an investment.
