Blog

The Hidden Costs of Building an In-House ESP

Building an in-house email platform can be tempting, but the hidden costs add up really quickly.
Dec 18, 2018
Nick Ziech-Lopez

When sending enterprise email at a large scale, a common obstacle that many B2C martech / marketing operations groups face is the struggle of moving and using their data. We’ve spoken about this topic in the past, and we’ve written about it too — but that’s because there are so many hurdles caused by the data movement necessary with a traditional marketing cloud ESP. Whether it’s setting up nightly copy and replication jobs to send your customer data to your marketing cloud ESP, fitting your data to their stringent data model, handling PII data security, or finding a way to bring your email data back down to your environment, using traditional ESPs at a massive scale can cause headaches for most marketing ops groups.

When these issues eventually become too costly and time-consuming (which is happening more frequently than ever), marketing ops is faced with a challenging dilemma:

  • Build a complete marketing cross-channel messaging solution in-house by creating a solution from the ground up or, more likely, utilizing a cloud-based email sending engine and custom building the rest of the required capabilities
  • Buy a production-ready on-premises cross-channel messaging solution that can integrate with your data

Both solutions can solve a company’s data and message-delivery problems. But, while there’s a lot to like on paper about the “Build” option, there are a lot of cost and time factors that are easy to overlook for a typical company that hasn’t built and maintained an ESP before.

Building an in-house ESP

The first variable to consider when thinking about building your own ESP is both the initial challenge and the ongoing work involved in developing the platform your marketing team will use for creating email content.

There are numerous components that need to be built into any platform just to start you on the same level as your competitors. Those will take time, and technology advances quickly; your needs and the standards of the market might even be different by the time you’ve completed building than they were when you began. But if you were planning to build an ESP platform today, these are the minimum structural requirements:

  • Audience Creation – It’s essential that marketing teams are able to create lists and dynamic audiences quickly and easily to target the right recipients with messaging. You’ll need a solution that allows them to do this quickly and easily.
  • Content Assembly – Marketers need simplicity and ease in a user interface, but your tech team also needs back-end access to deal with challenges. How will marketers be able to assemble and test message content?
  • Campaign Orchestration – Multi-channel and cross-channel journeys are an important part of competing with today’s customers. You’ll need a tool that allows marketers to build complex, dynamic, multi-step, cross-channel programs.
  • APIs for Automated and Behavioral Triggers – You need to set up and maintain APIs in order to send various automated and behavior-triggered campaigns from the website and other channels.
  • Message Rendering and Delivery Infrastructure (plus maintaining deliverability) – Processes will need to be set-up for message rendering and delivery. Ideally, you’d place a dynamic combination of audience and conditional content elements into specific message templates for email, push, SMS delivery across globally distributed endpoints.
  • User-level Engagement Tracking – You’ll need a way to track real-time engagement data on a per recipient basis. When did a user open, click, or take an action on a message? Did they receive the message at all? Did the user unsubscribe?
  • Campaign and Program Analytics – A robust analytics suite is key to understanding how campaigns are performing. You’ll need mechanisms in place so that marketing teams can create reports and dashboards that track everything that’s going on in each messaging channel.

And that’s just to get off the ground, not including many of the nice-to-haves that your competitors will have from their ESPs, or anything that the future market will require. Your team not only has to maintain all of these components, keeping them functional, but also has to develop new technologies as the competitive landscape evolves. It’s an expensive, time-consuming proposition that few companies are set up to be able to tackle.

Your team not only has to maintain all of these components, keeping them functional, but also has to develop new technologies as the competitive landscape evolves.

Message Rendering and Delivery Cloud Infrastructure — Special Considerations

Marketing operations groups in need of a more tailored solution to their needs will oftentimes consider using an API-driven cloud email delivery platform to send their email, seeking the promise of easier integration and greater flexibility. These platforms, such as SendGrid and Mailgun, tout the ease of getting up and running and will even give you free examples on their landing page of how to send sample emails right now. However, the ‘ultimate flexibility’ of these tools comes at a cost: the need to constantly maintain an email marketing cross-channel campaign management solution.

On the surface, this may seem easy — the example is right on the landing page! Let’s imagine you’re facing this decision right now, and take a look at a few realistic examples where building your solution may prove to be more difficult.

Because these tools have limited personalization or substitution capabilities (an issue many developers struggle with), you’ll probably have to build your own personalization solution. This could seem to be no problem, as there are many open-source platforms to build on (e.g. handlebars or FreeMarker) and you might even be able to build your personalization engine — but email is constantly changing. You’ll need to continually maintain your solution to include support for innovations such as:

  • New email clients’ recent support for Google AMP script
  • Any updated approaches to Universal Linking, which could change with each update to iOS or Android
  • Any changes to ensure that your messages are rendering well in the hundreds of email clients your users will view your emails in

Supporting an entire ESP solution is a very labor-intensive process.

Of course, this will take an entire team (or teams) to support. And speaking of teams: supporting an entire ESP solution is a very labor-intensive process. In addition to the simple personalization example above, you’ll need a person or set of people to:

  • Constantly test your solution for quality. With all of the constant changes happening in the email and cross-channel messaging space, your solution is going to have to change quickly to keep up. This means your solution is at risk of breaking down without manpower to test that it is production-ready
  • Monitor your IP reputation and deliverability. While many of these cloud-based solutions will offer guidance on deliverability, you are ultimately responsible for ensuring your messages will get to the inbox. Do you have deliverability expertise on-hand? How many IPs or domains will you be maintaining? These are key considerations when attempting to manage reputation in-house.
  • Help your internal users understand how to use the tool. How will users know how to create email templates, or send email? It may seem small, but documentation plays a key part in maintaining your solution

Keeping track of these teams and capabilities can be logistically difficult and very expensive. Which leads me to my next point:

  • An in-house solution will have a large amount of personnel and infrastructure costs. At first glance, the CPM may appear much lower with an API-based sender or SMTP relay — but that’s just the cost to send the email. I’ve previously written how much more expensive these solutions can be from an infrastructure and sending perspective, but that doesn’t even begin to consider the server and hosting costs the entire solution may include. Add on to that the cost of people I mentioned above, and the expense begins to balloon past the simple per-email cost these tools advertise.

This is not to say that in-house tools aren’t a good option for anybody, or that they are never a viable solution to an operational email problem. For instance, building your own solution might work for organizations that:

  • Have unique, or not well-understood email or cross-channel messaging needs, that could not be solved by any third-party ESP platform provider
  • Have a pre-existing large and skilled software development workforce that could easily provide the required skills (estimated at 10 engineers dedicated to ESP custom development)
  • Are comfortable with internal one-time and ongoing costs of ESP custom development (10 engineers x $200k fully loaded costs = $2m a year)
  • Accept the opportunity costs of investing internal resources on ESP customer development relative to other internal investment options
  • Are comfortable with the risks associated with custom development in terms of schedule delays and time to market
  • View ESP platform technology as core asset and intellectual property investment
  • Have unique email experience or partnerships that could be lent to building your own solution

It’s important for your Marketing and I.T. teams to take a realistic look at the full costs and investment it will take to essentially become your own ESP.

If your company meets all those criteria, “Build” could be a viable option if it’s something you’re considering. Otherwise, it’s important for your Marketing and I.T. teams to take a realistic look at the full costs and investment it will take to essentially become your own ESP for an indefinite period of time.

While the idea of having “free” email with a custom build can seem tempting, an ill-considered attempt at “Build” could cost your company millions in diverted resources, wasted efforts, and customers lost to competitors. Be sure you understand exactly what’s involved before you head down that road. And if, like the vast majority of companies, you decide “Buy” is your more prudent option, be diligent in finding one that solves your most pressing problems today. Prioritizing your needs will greatly help you find the ESP that’s right for your company.

About the Author

Nick Ziech-Lopez

Nick is the Senior Director of Product Marketing at MessageGears. He applies his background in engineering and data analytics to organizing his product backlog, understanding user experience, and obsessing over the Chicago Cubs.