Stagnation and comfort. That’s what I see in most marketing and sales departments regarding analytics. The usual answer to the question of what type of analysis and analytics stack you are using the answer is always the same:

  1. Track events
  2. Track sales
  3. Track CAC
  4. “I DON’T SEE THE NEED FOR ANYTHING ELSE”

‍

And these departments do it all with a combination of Google Analytics, Segment, Mix Pannel, Amplitude, DMPs and CDPs.

‍

Hint: If you are still using CDPs and DMPs you must definitely continue reading. 

‍

Marketing analytics and tools haven’t really evolved in the past 10 years since well, the year that companies like Segment and Mixpanel were born. 

‍

Let’s start with demystifying the CDP. CDPs are an added layer to the data stack that reduce flexibility. CDPs cannot serve as your single source of truth. Their “cookie-cutter” data models and reliance on tracking standard user events prevent you from properly representing their business-specific data models like products, groups, coupons, artists, etc. 

‍

Take for example Segment’s Personas. A person can only be attached to one account when in reality businesses are much more complex than that. People can live in different accounts. Take a business like Amazon for instance where a single person can belong to multiple accounts and have multiple specifics like coupons, discounts and so on. That’s one of the reasons data analysts and Data Scientists are using the Database approach where they can easily query the database and create the analysis that best fits according to their needs. That said, CDPs tend to disappear in the near future making the DB the new CDP. Building one source of truth is hard, so why build two? As data teams take form and organize, control and put in place data governance protocols at the enterprise level, the room for CDPs is smaller and smaller.

‍

Let’s jump to DMPs. Personally I’ve always seen DMPs as the worst investment ever, in part because everytime that a rigorous analysis is made on fraud, DMPs are the major source of it. Clicks are usually originated by bots. But that apart, DMPs are fed with what we call third party cookies and guess what, companies like Apple and Google had committed to eliminate them. That said, all the companies that have relied on DMPs must migrate to first party data, there’s no way around it no matter what DMPs are telling you and if we come full circle the best way to centralize your data is via your Database, Warehouse or Datalake.

‍

That taken out of the way let’s discuss the analytics part. It is understandable that marketing and sales want to analyze the basic and traditional metrics. Of course it is a must, but what’s not understandable is not wanting to look ahead with all the benefits that new technologies like AI can bring. The benefits include:

  1. cost reduction
  2. Increased efficiency 
  3. Collaboration
  4. Digital transformation 

‍

Cost reduction: AI has proven that it can reduce the Cost of Client Acquisition or CAC. In several of our clients the savings have been between 50% and 70%. How? By correctly measuring full attribution and then running Machine Learning Models to understand what, where and who to target.

‍

Increased efficiency: From buyer personas to hyper-personalization Machine Learning can truly make a difference in how you target your customers.

‍

Collaboration: By having a stack where the source of truth is the same as the one the data team uses and by using the same techniques and languages the whole team uses, collaboration just spreads like wildfire.

‍

Digital transformation: By adopting AI you are helping the company enter the next stage, you push the team to learn and bend the boundaries.

‍

Marketing teams need to start thinking deeper in their data operations and how they can start aligning with the company data strategy without losing its independence and flexibility.

‍

 Let’s go through a couple of examples:

‍

Buyer personas:

The first example comes from a very well known brand with thousands of points of sale. This company has specific information on their users like age, sex, purchases, where they buy, what promotions they like among many more. Marketing was using the same tools that most clients use and actually where using some predictive analytics. Problem is the predictive analytics used where pre defined standard analysis like RFM and recommended products. Other than that the analysis was always the same: Which users are buying what and where in general terms. What are my most visited domains and what are my sales per point of sales.

‍

The problem here starts with the tools, current tools are not helping this type of companies look ahead because they are restricted to an input and output. The team couldn't actually see beyond it. So when we approached them we helped them craft a solution where we built buyer personas with Machine learning based on many attributes. Actually more than 7 attributes. But not only that, we went on to recommend products, promotions and look and feel for each client. What today is called hyper personalization. Most interestingly we were able to work in conjunction with them to define the parameters of the model based on their understanding of their business.

‍

All of this raised a challenge, how would marketing change their operations to deal with this influx of new information? Isn’t that wonderful? That an initiative like this forces to redefine the operations of a whole division within a company?

‍

Point is, the tools used today are limiting companies because the standard is the basics. Marketing and Sales need to think and prepare for what's next. Yes we need to do the basics now but how do we prepare the organization and the team to do predictive analytics, quickly and easily?

‍

Datagran spent only 1 month with this client. By the end the client had a working model in production to test with its user base.

‍

10x customer experience:

One of the rules of thumb in the Startup world is whatever you do, do it 10x better. It is the only way to win. 

For example, let's take a marketplace company where their main objective is to connect users with distributors. One of the main analytics tasks in a Marketplace is to understand how many products each user buys from a distributor, what are the sales and when did they happen. The thing is that nobody within the company thought that the experience could be upgraded in days. We worked with this client to create a model that based on purchases, inventory, among other things it could recommend the products for the user but not only that, based on past purchases it would pre fill the shopping cart for the next order. We built this product with them in 30 days. This wouldn’t have been possible with any of the analytics tools currently used. Why? This are the things you would need to put together such a project:
1. System that can ingest data

2. Upload ML models

3. Re train models

4. Scale machines

5. Scheduler

6. Apis to connect to the client app

7. Data Engineer, DevOps, DataOps

‍

Not many companies have the time, resources and talent to put this together. Marketing and Sales departments have the tools today to pursue these types of projects. Yes, they need to solve the and visualize the metrics that are needed today, but they need to choose the right tools that will build the organization of tomorrow. There are no more excuses, stagnation is not an option, leaders just need the curiosity to serve their clients better and the tools are there ready to provide what’s needed to go the extra mile.

‍

Are you a next generation marketing or sales person?

‍