Should you use your data warehouse as your CDP?

1 year ago 74

The advent of cloud-based data warehouses (DWHs) has brought simpler deployment, greater standard and amended show to a increasing acceptable of data-driven usage cases. DWHs person go much prevalent successful endeavor tech stacks, including martech stacks. 

Inevitably, this begs the question: should you employment your existing DWH arsenic a customer information platform (CDP)? After all, erstwhile you re-use an existing constituent successful your stack, you tin prevention resources and debar caller risks.

But the communicative isn’t truthful simple, and aggregate imaginable plan patterns await. Ultimately, there’s a lawsuit for and against utilizing your DWH arsenic a CDP. Let’s excavation deeper.

DWH arsenic a CDP whitethorn not beryllium close for you

There are respective inherent problems with utilizing a DWH arsenic a CDP. The archetypal is obvious: not each organizations person a DWH successful place. Sometimes, an endeavor DWH squad does not person the clip oregon resources to enactment customer-centered usage cases. Other enterprises efficaciously deploy a CDP arsenic a quasi-data warehouse. (Not each CDPs tin bash this, but you get the point.)

Let’s accidental you person astir oregon each your lawsuit information successful a DWH. The occupation for many, if not most, enterprises is that the information isn’t accessible successful a marketer-friendly way. Typically, an endeavor DWH is constructed to enactment analytics usage cases, not activation usage cases. This affects however the information is labeled, managed, related and governed internally. 

Recall that a DWH is fundamentally for retention and computing, which means information is stored successful database tables with file names arsenic attributes. You past constitute analyzable SQL statements to entree that data. It’s unrealistic for your marketers to retrieve array and file names earlier they tin make segments for activation. Or successful different words, DWHs typically don’t enactment marketer self-service arsenic astir CDPs do. 

This besides touches connected a broader structural issue. DWHs aren’t typically designed to enactment real-time selling usage cases that galore CDPs target. It tin execute speedy calculations, and you tin docket ingestion and processing to transpire astatine predominant intervals, but it is inactive not real-time. Similarly, with immoderate exceptions, a DWH doesn’t privation to enactment disconnected earthy data, whereas marketers often privation to employment earthy information (typically events) to trigger definite activations.

Finally, retrieve that information and the quality to entree it don’t maketh a CDP. Most CDPs connection immoderate subset of further capabilities you won’t find successful a DWH, specified as:

  • Event subsystem with triggering.
  • Anonymous identity resolution.
  • Marketer-friendly interface for segmentation.
  • Segment activation profiles with connectors.
  • Potentially testing, personalization and proposal services.

A DWH unsocial volition not supply these capabilities, truthful you volition request to root these elsewhere. Of course, DWH vendors person sizable spouse marketplaces. You tin find galore alternatives, but they’re not autochthonal and volition necessitate integration and enactment effort. 

Not surprisingly, then, there’s a batch of chatter astir “composable CDPs” and the imaginable relation of a DWH successful that context. I’ve argued antecedently that composability is a spectrum, and you commencement losing benefits beyond a definite point. 

Having issued each these caveats, a DWH tin play a relation arsenic portion of a lawsuit information stack, including:

  • Doing distant with a CDP by activating straight from the DWH. 
  • Using the DWH arsenic a quasi-CDP with a reverse ETL platform.
  • Coexisting with a CDP.

Let’s look astatine these 3 plan patterns.

1. Connecting selling platforms straight to your DWH

This is possibly the astir utmost lawsuit I critiqued above, but immoderate enterprises person made this work, particularly successful the pre-CDP epoch and platforms (like Snowflake with its wide ecosystem) are looking to effort to lick this.

The thought present is that your engagement level straight connects to push-pull information with a DWH. Many mature email and selling automation platforms are natively wired to bash this, albeit typically via batch push. Your marketers past usage the messaging level to make segments and nonstop messages to those segments successful the lawsuit of outbound marketing.

Marketing platforms straight  ingesting from DWHMarketing platforms straight ingesting from DWH

Imagine you had different selling oregon engagement platform, a personalized website oregon ecommerce platform. Again you gully information from DWH, past employment the web exertion level to make different acceptable of segments for much targeted engagement.

Do you spot the occupation yet? There are 2 sets of segmentation interfaces already. What happens if you had 10 selling platforms? 20? You volition support creating segments everywhere, truthful your omnichannel committedness disappears. 

Finally, what if you had to adhd different selling level that did not enactment nonstop ingestion from a DWH?

This attack solves respective problems with the archetypal signifier above. Notably, it allows (in theory) a non-DWH specializer to make cosmopolitan segments virtually atop the DWH and activate aggregate platforms. With translation and a amended connector framework, you tin use antithetic statement mappings and marketer-friendly information structures to antithetic endpoints.

Here’s however it works. Reverse ETL platforms propulsion information from the DWH and nonstop it to selling platforms aft immoderate transformation. You tin execute aggregate transformations and nonstop that information to respective destinations simultaneously. You tin adjacent automate it and person exports tally regularly astatine a predefined schedule.

Reverse-ETL tools tin  enactment  arsenic  an intermediary furniture   for modeling and activationReverse-ETL tools tin enactment arsenic an intermediary furniture for modeling and activation

But a transcript of that information (or a subset of it) is really copied implicit to people platforms, truthful you truly don’t person conscionable a azygous transcript of data. Since the reverse-ETL level does not person a transcript of data, your required segments oregon audiences are ever generated astatine query clip (typically successful batches). Then you export them implicit to destinations.

This is not a suitable attack if you privation to person real-time triggers oregon always-on campaigns based connected events. Sure, you tin automate your exports astatine precocious frequency, but that’s not real-time. As you summation your export frequency, your costs volition exponentially increase.

Also, portion reverse-ETL tools supply a segmentation interface, they thin to beryllium much method and DataOps-focused alternatively than MOps-focused. Before declaring this a “business-friendly” solution suitable for marketer self-service, you indispensable trial it carefully.

3. DWH co-exists with CDP

Your endeavor DWH serves arsenic a lawsuit information infrastructure furniture that supplies information to your CDP (among different endpoints). Many, if not most, CDPs present connection immoderate capabilities to sync from DWH platforms, notably Snowflake.

CDP and DWH tin  co-existCDP and DWH tin co-exist

There are variations successful however these CDPs tin co-exist with DWH. Most CDPs sync and duplicate information into their repository, whereas others (including reverse-ETL vendors) don’t marque a copy. However, determination could beryllium trade-offs you request to see earlier finalizing what works for you.

In general, we thin to spot larger enterprises preferring this plan pattern, albeit with wide variance astir wherever specified captious services arsenic lawsuit individuality solution yet reside. 

Dig deeper: Where should a CDP acceptable successful your martech stack?

Wrap-up

DWH platforms play progressively indispensable roles successful martech stacks. However, you proceed to person aggregate architectural choices astir which services you render wrong your information ecosystem.

I deliberation it’s premature to regularisation retired CDPs successful your future. Each signifier has its trade-offs to support successful caput portion evaluating your options. 


Get MarTech! Daily. Free. In your inbox.


Opinions expressed successful this nonfiction are those of the impermanent writer and not needfully MarTech. Staff authors are listed here.

Read Entire Article