DMPs in adtech handle anonymous data, where customers' information cannot be traced back to their identities based on the use of randomly generated cookie IDs. On the other hand, CDPs store known information about customers, such as their contact information, purchase history, sentiment, and membership tier. Anyone with access to your CDP can easily dig out anything about your business - financial health, core user segments, future growth potential, even your weaknesses.

This valuable and sensitive data, if left in the hands of a third-party vendor, can expose a company's core business intelligence to unauthorized parties. For example, your competitors may dig out secret sauces from your vendor service team. Your vendor may study your data for future upselling and renewal strategy. Hackers may your data, exposing you to huge regulatory fines.

For mid to large-size enterprises, it may be more beneficial to have an in-house CDP rather than relying on a SaaS-based solution. This can lead to cost savings, easier organization transformations informed by data, and better control over optimizations.

However, every company's situation is unique, and careful consideration and evaluation should be made before making a decision.

I can help figure out the best strategy for you.

(1) Overwhelming Analytics:
Employees are becoming overwhelmed by an excess of dashboards and a lack of clear purpose. From my experience, dashboards follow a long-tail distribution, with only a small percentage (less than 5%) being regularly monitored by stakeholders. The rest become ineffective due to outdated data, redundant content, or a lack of meaningful context.

(2) Analytics without Business Goals:
Visualizations can be visually appealing, such as Uber's open-source visualization tool https://kepler.gl/. However, as enterprises, we must aim for more than just aesthetics. Too many employees are creating dashboards simply because their superiors want to see them, without considering their purpose.

The ultimate goal of analytics should be to help the enterprise achieve its business objectives. Analytics is an iterative process that involves understanding the current state, identifying problem areas, proposing solutions, implementing changes, evaluating results, and repeating as necessary.

Therefore, if a dashboard does not lead to proactive monitoring and action, it may be more beneficial to discard it.

Focus on the outcome, not just the means of presentation.

In the past, data analysis within organizations often involved a limited number of static tables that could be managed by a single person. However, this traditional approach is no longer sufficient in today's rapidly changing data landscape. This is because customer data has become real-time, high-volume, and unstructured, and companies are utilizing multiple SaaS vendors with varying data formats.

To overcome these challenges, enterprises require robust ELT tools that can effectively manage the gap between data ingestion and data activation.

Here are some of the common challenges in ELT:

- Merging data from various sources and understanding the meaning of each data column.
- Resolving conflicting data fields and determining the reliability of data sources.
- Combining real-time data from APIs with offline batch data from large files or databases.
- Catering to the needs of data analysts and addressing last-minute requests for new columns or formats.

These are just a few examples of the numerous difficulties organizations face in ELT.

Unfortunately, many companies initially believe they can build an in-house data platform, only to later realize that they are ensnared in the ELT rabbit hole.

However, I bring extensive experience in handling ELT and data readiness for multiple major data platforms, and I am confident I can assist you in overcoming these challenges.