My First CDP: Why I chose a CDP as my ESP replacement


Every once in a while, during our careers, we get to do something truly exciting. Launch a new product. Solve a business problem in a novel way. Build a new system. On rare occasions, we get to do all these things at once. This is one of those stories.

These situations present opportunities to create, to fulfill a business need and, most importantly, to learn something new and transformational that may launch our career in a new direction. I encountered one such situation in 2016 during the course of a Request for Proposal process in the search for a new ESP. As you can guess from the title of this post, I chose a CDP, then a brand new category of platforms of which I had never heard. Full disclosure -- the CDP offered a fully baked integration with several well-known and low-cost transactional email senders. So, while we still technically had an ESP, we shifted the bulk of the ESP budget to the CDP instead, which quickly became our day-to-day marketing operations hub.

As the marketing tech ecosystem has grown over the past decade, a multitude of new platforms offering a plethora of new capabilities have exploded onto the scene. This makes it increasingly difficult to find new platforms and vendors that do what you need and not what you don’t…and that don’t overlap with each other. With each new generation of martech platforms, new acronyms emerge describing the latest must-have capabilities. First ESPs burst onto the scene in the early 00s, then CRMs, then DMPs. More recently, CDPs arrived with great fanfare and have attracted quite a haul of angel investor dollars over the past few years.

Just last year, in a June 2019 article titled “Know your CDPs from your CRMs and DMPs,” reported that the “CDP (customer data platform) has emerged, almost overnight, as yet another must-have TLA (or “three letter acronym”) that every marketing team simply cannot live without.” The article went on to cite CDP industry revenue hitting $740 million in 2018 and projected that revenue would exceed $1 billion in 2019.

The article laid out 5 key characteristics that marketers should look for in a CDP platform, including:

  • Putting data integration into the hands of marketers and enabling data to be integrated easily from disparate sources, creating a single view of the customer
  • Easy-to-use and powerful segmentation capabilities that can enable data to be used in ways that have become difficult with traditional filter and query creation tools
  • Engagement / real-time trigger programs via API integrations with ESPs and other messaging channel provider platforms
  • Data orchestration across multiple platforms, databases, other martech systems and channel communication platforms (email, push, SMS, etc)
  • The transformation of data across channels into reportable and actionable events

At the time this article was written in 2019, the CDP industry was already several years old. As I mentioned, I first discovered CDPs in 2016, having stumbled upon the emerging platforms as a result of a cold-email acquisition campaign. It just so happened to coincide with the email RFP process I had just begun. At the end of that RFP process, I won’t be giving much away at this point by saying that I selected the CDP over many well-established and some smaller, nimble ESPs.

A lot has been written about what comprises a CDP and how it is different from an ESP, a CRM, a DMP, etc. What was interesting to me in this selection process is what made the CDP so similar to an ESP, in certain respects, that it could be included as a natural fit for our ESP search. The linking factor in my case was a data integration challenge. In this post, I will discuss how I found my first CDP, why I chose that solution as an ESP replacement, and the benefits we derived from using a CDP as our marketing operations hub to manage our email and push campaigns.

In 2016, I took a job heading the retention marketing practice of a consumer-focused cloud software and services subscription business. The company had an urgent deliverability problem at the moment, so my experience in deliverability, customer lifecycle retention and loyalty programs landed me the job. Over the course of the first few months, we navigated a successful recovery plan to our deliverability challenges.

We first fixed some technical misconfigurations on our mailing IP and subdomains. We setup our SPF, DKIM and DMARC records. Then, we focused on mailing our most engaged customers, separated our programs by transactional and marketing IPs and religiously tested subject lines. Our sender reputation, inboxing rates, open rates and engagement rates dramatically improved with discipline, measurement, optimization and, most importantly, with time.

We saw the most dramatic results with our welcome email series which was sent to new subscribers. It contained useful information for how to get started using our software and services. Other programs plateaued during the process and remained stuck at stubbornly low engagement rates. After reaching out to our Email Service Provider account manager and then engaging with the company’s deliverability professional services team, we were displeased with the level of competence and service provided by that company. We decided to search for a new Email Service Provider.

I decided, if we were going to put together an RFP and switch providers, I wanted to do it right. So, we first did a gap analysis. What tools and data did we have access to today and where do we want to be within one year? The current setup with our existing ESP was very list-based with a once-a-day bulk load data update from our customer database. Our email database contained 20 or so fields which we could use to create segments or personalize emails by first name.

Our business had fundamentally changed since my predecessors set up this infrastructure. We ran a highly successful web application with tens-of-thousands of paying subscribers and planned to launch two new apps in the coming months. With our simplistic email database, we had access to none of the data generated from usage of our products or website, only highly aggregated subscription statuses, status change dates, signup dates and the like.

So, we first sought to find and integrate an analytics vendor who could tag our website, our behind-the-paywall web application, as well as the iOS and Android versions of our two new apps.

As part of the rollout of the new apps, we planned to introduce a new push messaging channel to communicate with subscribers during their onboarding process. So, we needed to add to our Email Platform Vendor RFP a requirement to support push messaging – preferably both in-app push as well as web push. This didn’t seem like a big deal, as many of the traditional ESPs were transforming themselves into multi-channel marketing clouds.

We wanted to connect our customer billing system to the marketing database. Having access to customer billing information and history would allow us to tally up the spend of each customer over time and give our marketing program full access to the “m” in RFM (recency, frequency and monetary value). This capability would enable us to target high value and high loyalty cohorts with specific messaging. This requirement, along with the integration of the web and app analytics event data, meant that the new email vendor had to support relational and transactional data structures within customer profile data.

Next, we planned to integrate our ZenDesk customer service data with the marketing database. This would enable us to message customers with specific types of customer service issues, messaging geared to address their concern or technical support category.

Finally, we wanted to integrate all the marketing touchpoints and response data into Tableau dashboards.

These requirements formed quickly around the new ESP search and we soon issued the RFP with the 5 key things we were looking for:

  • Integrate our Analytics vendor of choice and pull in events from our website, web application, iOS and Android apps
  • Support email and push messaging channels
  • Integrate customer value and billing history
  • Integrate with ZenDesk
  • Integrate with Tableau for reporting dashboard creation

If successful, this new “marketing stack” would provide enormous organizational value to the company and extend our core marketing technology capabilities in strategic new directions.

We hired a website optimization expert to lead the search for the analytics vendor and I mapped out the ESP search. We came together as needed to make sure we found a solution that would 100% work together.

I began the process with interviews of six leading ESP vendors, representing a wide range of established platforms and newer, more nimble players. That initial list included Epsilon, Salesforce, IBM, Oracle, QuickPivot and Cordial. One vendor politely exited early from the process, as they required a minimum annual commitment that was completely out of our league as a small tech startup with big money but bigger dreams.

A little further along, one of the established marketing clouds informed us that they required us to hire a 3rd party agency to perform base-level system setup and integration tasks. This added significant incremental cost to their proposal. Duly noted.

Fortunately, I had included some smaller, newer, nimbler and hungrier companies in my process. These platforms showed flashy new applications with lots of functionality and U-eye candy. These bells and whistles along with more flexible modes of data integration appealed to the creative “startuppy” types that comprised our marketing, technology and executive teams.

The final step was for each company to do a demo of their product based around the requirements and then to submit a pricing proposal for a 1 year term. The demos went well and near the end we had some clear favorites and a decent start on a comparison scorecard. That left the pricing proposal phase. Here’s how that went.

Pricing for the centralized marketing database we envisioned that could host lots of event data from multiple sources turned out to be a major cost-driver. Proposals ranged from $60k to more than $100k to setup and customize the database, then steep monthly maintenance fees ongoing. Not only that, once the database was set up, we would have to write APIs to extract the data from the disparate analytics, billing, and customer service data sources and write it to the new email vendor’s database. That effort alone would require significant technical resource to which only our CTO could commit.

Suddenly, with huge up-front costs and technical resources that would be required to perform the integrations, my little ESP replacement project turned into a costly and time-consuming endeavor for our small startup company. Then something happened that completely changed the course of the project.

In the very early stages of the RFP process, I remembered getting a cold email from a guy whose name sounded like a spice and who worked for a company with “Data” in its name. It was a plain-text email. At first, I ignored it. Then, I remembered ignoring it a second time. I think it was the third time receiving an email from a guy named Basil that I clicked the website link in his email signature. I think I had started to realize the costs for what we ultimately wanted to do might be prohibitive. I liked what I saw.

Simon Data’s website featured retention marketing, partner integrations and customer data management across disparate sources as their calling cards. After a phone consultation with Basil, who was smart, consultative, energetic (but not overbearing) and personable, I invited them into our ESP RFP process.

Simon Data boasted a fully baked bidirectional out of the box API integration with SendGrid, along with several other email platforms who seemed to specialize in transactional messaging. SendGrid just so happened to be integrated into our web application for certain triggers that had been coded by our engineers. So, we already had a functioning SendGrid account that was sending product-triggered emails.

Moreover, Simon offered integrations with Heap Analytics, who had recently made our short-list of analytics vendors, could ingest ZenDesk data and also offered integrations with many push messaging platforms.

SendGrid, in particular, played a heavy role in easing concerns about how our email program would transition. It would be the first-mover and needed to make a smooth transition. Since we already had a SendGrid account, I could easily get my head around how the messaging could be triggered from Simon’s platform and how the reporting would work on the backend. In a way, it made sense to move our marketing messages over to SendGrid and unify marketing emails with transactional / product-triggered emails on one platform. This had the bonus side-effect of generating cost savings vs. our existing ESP, so more budget could be shifted to the CDP.

The final deciding factor was Simon’s method of pulling all our data from these disparate systems into one place and making it actionable. No APIs had to be written (hooray!)…only queries. We had to dedicate very little DBA and technical resource from within our company to make this integration happen.

Within 4 weeks, we had most of our data pulling into the Simon Data system, updating multiple times per day and we were sending email and seeing across the board improvements in deliverability and engagement rates. We could easily (and affordably) add new data from new systems with a simple integration and a query.  Of course, the limiting factor here is sometimes which other platforms that particular CDP integrates with most easily.

We discovered many significant benefits from leveraging a CDP as our go-to marketing operations hub.

  • Data unification – Gaining access to all our data across products, sites, apps and systems in one unified marketing-controlled environment was revolutionary.
  • Data orchestration – Marketing gained control of the data orchestration between our disparate systems. We had complete control over how the data was queried, how it was transformed for use in marketing segments and key metrics, and the timing at which the data pipeline refreshed.
  • Direct integration to all native data sources
  • Multiple keys to join data – This was something I didn’t think about at first. Most of our internal information was keyed on a user ID but one of our most lucrative email programs was our abandoned cart series. When someone abandons the signup process, you often only have an email but no user ID. Simon could join data by either or both, which created a very flexible environment for joining all our data together seamlessly.
  • API integration with multi-channel messaging platforms
  • Leveraging a single platform for managing email and push channels as well as marketing ad hoc and transactional / product-triggered messaging meant we only had to automate in one place.
  • Data Playground – A wise person recently said that many CDPs often meet the 5 criteria outlined above and then do something else as well, a differentiating factor. Simon Data’s dashboards delivered the something else in this case – something we had not explicitly looked for but which ended up being transformational for our marketing team.

Email teams have evolved over the past several years. By virtue of owning the email database, which has increasingly become the customer database, email marketers inherit stewardship of what is arguably any company’s crown jewel – its customer data. Increasingly, email marketers manage multiple channels of communication with customers and prospects and require complex aggregations of customer data that comprise the single view of the customer. As companies have realized the value of customer data, they have taken great strides to unify customer data. Marketing magic happens when you can make that unified customer data actionable.

Finding a new marketing technology platform that does all the things you want and none of the things you don’t – and that you can afford within your budget – is never easy. There are literally thousands of potential choices, many of which overlap in specific ways.

The ideal solution for you may not fit neatly into one category or platform. Yet, when you introduce multiple platforms into a marketing stack, you risk some degree of feature overlap. Within a thriving martech marketplace, feature overlap is nearly impossible to avoid.  The trick is to engineer strategic overlap to either create a failover option should one platform experience downtime…or to leverage the known overlap as a point of pricing negotiation.

I predict big changes ahead in the way marketing technology platforms are categorized and the blends of features and modes of integration offered. For example, traditional lines between adtech and martech ecosystems are blurring, with the impending death of the third-party cookie.

I always thought of adtech and martech as two separate things – complementary but with distinct purposes. Adtech platforms focused on acquisition and centered around leveraging 3rd party cookie networks. AdTech platforms worked to bolt on as much detail as possible to an otherwise anonymous user profile based on anonymized searches, purchases and browse behavior…all for the explicit purpose of serving contextually relevant ads and retargeting campaigns.

Martech platforms, meanwhile, focused on customer retention and centered around managing customer data and targeting campaigns to known customers and prospects. This is the world to which I am more attuned. But with Google announcing last year that Chrome would kill off third party cookies in 2022, I wondered how things will evolve (or converge) in the adtech and martech ecosystems over the next several years.

This change could make irrelevant (or a lot less important) the classic distinction between adtech and martech systems – that being anonymous vs. known user profiles. If both are going to focus largely on known profiles and nurturing relationships with customers to provide value in exchange for their data, will we see tectonic shifts as these two ecosystems come crashing together for the first time?

Obviously, by selecting a CDP as our marketing operations hub instead of the ESP for which we had been looking, we acknowledged some blurring of the lines between the two. Could the distinction between DMPs and CDPs also effectively disappear and lead to consolidation across multiple categories (CDMP, anyone?)? What mergers and acquisitions make sense in a third-party cookieless world? And how can we find the best martech solutions for capturing and managing our customer data or sending our messages in a constantly evolving marketplace?

It will never be easy. Sometimes the marketing technology solutions you find may not be the ones you started out looking for. Keep an open mind. Your search may take interesting twists and turns. Consider solutions that may fall into new categories, might not belong to a category of which you’ve ever heard, or might overlap in some ways with existing solutions. Twists of fate (or even cold email solicitations) can sometimes take you in new directions if you open your mind to solving problems in a novel way.

Related Posts