MarTech and data architecture for B2B Marketing Transformation

Having discussed the importance of using Predictive Analytics in your B2B Marketing, in this article, I will focus on explaining the data architecture needed in order to put your first use case (customer acquisition) in production. 

Your customer data is most likely scattered in a dozen different tools and databases; some data resides in your marketing automation tool, other in your CRM or ERP system, maybe on your DMP or Google Analytics account, even on excel spreadsheets and CSV/JSON files. Wouldn’t it be great if you could gather all this data in one central database which would act as a single source of truth for all your marketing, sales and customer data? In that way,  your data would be collected, managed and transformed into useful signals that you could later use in order to deliver personalized experiences to your audience, at scale.

MarTech Architecture & Data 

How would your marketing stack look like if you connect all of your B2B data into one central database? Well, something like this:

View fullsize

The idea is simple; you get all of your data sources and you connect them via an API to the central database (assumption made:  your Marketing Automation Software, your CRM & ERP systems are already integrated with each other). Once your data is gathered in the database, you can then choose to visualize the part that you are mostly interested in (depending on which department you belong to, for instance) using any Business Intelligence software . But your job is not done yet; you also have to act on your insights. Thus you pass the specific actions that you have decided into your Marketing Automation software which will act as an orchestrator of your activities. Nice & neat, right?

Technical Challenges

Building the architecture of such a project is definitely not a walk in the park. Here is a short list of what you should think beforehand:

• How many integrations will I need? Should I spend time on an integration or uploading a CSV file once per month will do the job? What is the associated cost in terms of time & resources?

• How often should the integrations run? How many records should I sync? Any limitations on the  API calls that I am allowed to make on a daily basis? 

• Which data should I sync & from which data sources?

• I have different formats for ‘date’ field in my CRM & Marketing Automation software. Which one should I keep? (of course this challenge will exist for a lot of your current fields)

• How do I ensure data security & governance?

• Do I have the resources that will maintain and support the integrations once the PoC (proof of concept) project is completed?

Building a solid infrastructure from scratch will require a substantial investment in terms of time and money. The good news is that there is another option; you can buy a SaaS solution, test your hypothesis and understand if that is the way forward for you. Hence, from a technical perspective, you can use existing real-time integrations and not worry about their maintenance  while also having a central massive database organized according to your business requirements. That is quite awesome!

From a business perspective, it would be prudent to define the use cases first, before you start building the architecture. If you know what you want to achieve, it is much easier then to solve the technical questions that will arise on your journey. Shall we start?

Your first use case: New  Customer Acquisition

How can we decide which customers to focus on, so that we avoid wasting resources on developing relations that ultimately will not create opportunities for sales, in a systematic way? We have to identify new customers that belong to the segments that we have already defined, while they are on the ‘ready-to-buy’ phase. A bit complicated, isn’t it? Let’s break that down.

Look-alike modeling is the answer to the first part of our challenge. Look-alike modeling is a process that identifies people who look and act just like our target audience. Hence, by analyzing our existing customers, we will identify some key characteristics & patterns that will enable us to spot the potential available customers. We will start by defining what is our ‘ideal existing customer profile’ & search for similar profiles out there. Some criteria that can be used: Revenue, profit margin, type of company, company size, expected revenue etc.

Real-time data from your Sales Prospecting tool will solve the second part of the puzzle. This software will be able to provide you with data that you can leverage to estimate when to approach prospects from a timing perspective. For instance, a company has recently opened their offices in your country, a company has just implemented a software that is complementary to yours etc.

Now, imagine how much better the quality of your prospects will be if you combine those two algorithms and attach a score on them which will indicate who is ready to buy. Thus your Sales team will only get those prospects that are really ‘Qualified’. Of course, you could also combine those two algorithms with the data from your website and improve the qualification even more. Then, once the Sales rep reaches out to the prospect, proper feedback will be stored in the CRM and thus circulating it back to the algorithms for further adjustments.

Bringing it all together, a central database (also known as Customer Data Platform or CDP) is a vital & much-needed marketing system that provides a single point of real-time control over customer data by unifying all types and sources. Although this could be a long implementation journey, you can use existing SaaS solutions to test the waters. Before doing so, define the use cases that make sense to your business. In my next articles, I will describe three more use cases (churn prediction, customer activation, cross/up sell) that are relevant to any B2B company. Stay tuned!