An introduction to building your own MarTech stack


In our previous article “MarTech stacks and where to start” we introduced the concept of MarTech and MarTech stacks. We reviewed the core challenges that this growing industry creates and what you, as a marketeer, consultant or product owner can do to overcome these challenges whilst building your very own MarTech stack.


In this article, we’ll take it a step further and look at (1) the different approaches you can adopt to building a MarTech stack and (2) review the main pro’s and con’s for each.



Different schools of thought


Historically, there have been two different schools of thought to building a MarTech stack:


1. The ‘monolith’ approach: buying your way into large integrated (marketing) suites. Essentially, betting on the vendor's philosophy, vision and capacity to develop their solutions over time. Examples of these are the Adobe Experience Cloud and Salesforce Marketing Cloud.


2. The ‘best-of-breed’ approach: selecting specific end-point solutions to address a specific marketing and business challenge, whilst keeping a holistic view vs. integration on the ‘stack’ you’re building.


Today, these approaches are still very much your best bet. Although, recent research (a 360 review of the integrated suite, Gartner 2019) has shown that more and more brands are moving towards and opting for the best-of-breed approach.


The driving factor behind this trend is the belief that if a brand is able to combine many specific end-point solutions and operate them effectively, they can create a competitive advantage. In short, best-of-breed stacks (when operated effectively) allow you to better understand, communicate with and serve your customers overall. We shared our thoughts on this in our previous article that can be found here.


More recently, and next to the two previous approaches, we’re seeing an increasing number of brands develop custom solutions. Often developed in-house through cloud instances. The reason being that there’s no solution (yet) that addresses their specific business needs. We’ll call this one the ‘custom’ approach:


3. The ‘custom’ approach: developing in-house solutions that are entirely tailored to your own business needs. These are usually more specific in scope and not scalable outside of your business context. E.g. leveraging a cloud instance to compute enhanced customer traits (CLV, churn, ... ) and sending this information to downstream solutions in the stack such as ESP’s and ad platforms.


Three different approaches that each craft their own way. This begs the question: “Which one is right for my business?”. In fact, we believe that all of these approaches can work for your business. Careful consideration of each approach’s pro’s and con’s will help you identify what direction to take. Let’s have a look.



The monolith approach


With this approach, you’re buying your way into large integrated (marketing) suites. Essentially, betting on the vendor's philosophy, vision and capacity to develop their solutions over time.

These vendors typically provide a range of solutions that integrate well as long as you stay within the boundaries of their own ‘stack’. This last part is key here. Point-in-case: try sending segmented analytics data from Adobe’s stack into a Google Ads Remarketing audience. You’ll find it’s an unnecessarily complex and painful process.


Customers are incentivized to pick and add-on solutions from the same vendor. The result of this is that customers face the choice of continuing down the road - and the inevitable vendor lock-in - or fully modernizing their entire MarTech stack.


Pro's

  • Single Vendor Management - you are interacting with a single vendor and thus should be able to count on more integrated support.

  • User onboarding - onboarding of users and usability perception ‘feels’ good (see above). No need to switch and ‘log-in’ into 10 different solutions, more centralised training opportunities, …

  • More seamless workflow across tools - Workflows, logic, and things such as naming conventions will usually be shared across the different solutions.

  • Solution integration (within suite) - onboarding of new solutions (from the same vendor) is easier because they integrate well and accessing the new solutions is literally only a few clicks away.


Con's

  • Risk of vendor ‘lock-in’ - The more solutions you add, the more you become reliant on one single vendor. By extension, your entire marketing organisation becomes dependent on the vision and evolution of this one particular vendor. Switching to a new vendor is guaranteed to be a cost and time intensive ordeal. Choose wisely.

  • Loss of features - They only integrate well within their own suite, leaving you empty handed whenever you want to integrate a solution that sits outside of those boundaries. You lose a lot of best-of-class features and capabilities.

  • Lower (budget) flexibility - If you want to evolve and build out your stack, you have little choice then to choose and commit to the solution provided by that same vendor. Cost structures can ramp up very quickly, since many of the solutions are priced separately as add-ons.

  • Slower innovation - the rate of innovation of these marketing suites is generally lower than the different end-point solutions. The former needs to evolve and drive innovation on all fronts, while the latter have a more focussed approach and can thus innovate faster.



The “best-of-breed” approach


Specialized players are offering an alternative for the Marketing Cloud fatigue: market oligopoly (Adobe, Salesforce, Oracle,...) which comes with difficult or limited integration, expensive pricing and inflexible data models.


The result is the creation of a “best of breed” approach where organisations select different technological components in order to build a unique stack tailored to their business needs.


Pro’s

  • Leverage best-in-class - you’re selecting solutions for their best-in-class features and integration capabilities. No need to compromise on features as you have an entire range of solutions to pick from. You’ll be able to get the most out of your MarTech stack when integrations are properly taken care of (see con’s).

  • Agility to react to new environments - the consumer needs are always changing and so is the MarTech industry. Almost on a weekly basis, new solutions are developed that try to address a new consumer need or challenge. Having the flexibility to test and adopt these solutions as they come is a powerful mechanism towards remaining at the edge of the industry and competitive.

  • Easier platform migration - your business changes, so do your (marketing) needs and challenges. As such, you should be able to change your MarTech solutions when needed. With a best-of-breed approach you are able to, relatively to the monolith approach, easily take out and migrate to a different solution altogether.


Con’s

  • Need for strong vision - you are interfacing with many different vendors, each with their own development and feature improvement roadmap. Juggling a bunch of these can quickly become overwhelming. Having a strong vision on how the different solutions and components integrate in a consistent manner is key.

  • Training & knowledge sharing - staying up to date with the latest of each solution is a challenge. Strong focus on always-on learning and knowledge sharing are keys to success.

  • Inconsistent integration patterns - Many end-point solutions have direct keypoint integration or come with strong API’s and webhooks allowing you to integrate across solutions. Nevertheless, data models and integration needs are often changing which makes for a complex exercise and require up front planning and assessment.

  • Complex workflows - activation workflows span across different solutions. E.g. Sending a centralised audience to Google Ads for remarketing purposes and linking to triggered emails will require you to interface with 3 tools (CDP, Google Ads, ESP). Setting up an infrastructure (and audience) blueprint are must-haves to keep a birds-eye view.



A third, more custom approach


More recently, a third approach is growing in ‘popularity’. This one’s for the more mature organisations out there: in-house building of custom apps or solutions that address a specific business use-case. E.g. computing enhanced user traits (CLV, Churn, …) to use in end-point solutions as mentioned in the introduction above.


While all of this does sound very exciting, it’s difficult to pull off and be effective in practice. On the flipside, doing this well can propel your marketing efforts and business to a higher and more profitable level. Our view is that these projects and custom ‘enhancements’ should be combined with either of the previous two approaches.


Pro’s:

  • Specific to your business context - the main reason for developing custom solutions. You are 100% in control and can direct the solution to address a specific business challenge or need. There will be vastly different solutions coming from a retail player vs. telco player.

  • Cheaper from a licensing point of view - you typically won’t have to pay for a SaaS license, bar some limited cloud storage costs if that’s used. Realistically, the main costs you’ll bear are developing resources, which can be expensive in the short but worth it in the long run.

  • No/little dependency to external parties - you are developing in-house. No solution vendor who you're dependent on, and unless you’re outsourcing development to an external agency, you’re 100% in control.

  • High agility to react to environments - you can develop your own solutions as your business context and needs evolve. Defining your own solution and feature roadmap gives you the control and agility to get an edge on your competition.


Con’s

  • Resources intensive - successfully developing custom solutions and stack enhancements demands quite some resources. You’ll be dependent on data architects or engineers to do the job. Alternatively, development can be outsourced to external partners, but that would make you lose some flexibility and kind of misses the point.

  • Requires very strong vision, structure & organisation - in order to successfully integrate custom solutions in your MarTech stack, you need to know what you’re doing. The key is that your own solutions not only integrate well, but also enhance whatever you have in place.

  • Onboarding is never standard - every solution you develop will have its own integration needs and profile. Furthermore, onboarding of your team on solution use will be different each time.



Our view


As a closing note, we believe that building a robust and integrated MarTech ‘foundation’ with out-of-the-box solutions, either from the same vendor or a selection of best-of-breeds should be your first focus. Custom solutions will set you up for rather incremental gains and advantages in the short run (but can compound in the long run).


Let’s not forget that, also here, the Pareto principle can be applied: in spirit, 80% of your business challenges can be addressed with the most basic solutions at hand. Backed by a strong vision, keep it simple at first, get the basics right, and slowly but surely add more advanced solutions to the mix. Only then, and only then, should developing custom solutions be considered.


Not sure where to start or want to know how Human37 can help? Reach out!