Off the Shelf, or Off to Work? Here are the questions to ask and aspects to consider when deciding whether to integrate with a ready-made 3rd party marketing technology solution or to build it in-house
The marketing technology space is complex, ever-changing, and, sometimes, full of hype. Many organizations struggle to find their ways in and around it, which new platform or service to purchase or integrate.
Traditionally, marketing organizations make these decisions on their own. They would create the martech roadmap and select vendors, relying on IT mainly for deployment and integration. But Gartner’s recent Marketing Technology Survey 2019 (subscription required), shows a change in this trend, with Marketing and IT navigating the confusing waters of the martech ocean closer together than ever before.
The increased collaboration between departments should lead to improved decision-making when evaluating martech ecosystems and choosing the right solution. But it also opens a new series of questions, highlighted by the dilemma of whether to buy or build the solutions they need.
A few weeks ago, I came across a fascinating interview by Scott Brinker from Chief Marketing Technologist with Lyft’s engineers who built a proprietary marketing automation platform tailored explicitly for ad-bidding automation. The interview got me thinking about what the guidelines for marketing teams approaching this question should be.
What to take into account
1. Resources – Building a marketing automation solution in-house is resource-intensive. For Lyft, it required the joint effort of the marketing, IT, engineering, and data teams. The core team consisted of 7 engineers and 4 scientists with 2 marketers embedded in the team to provide context to requirements and QA product versions.
On the other hand, buying a solution requires the investment of the same teams if not more, with the main difference being the amount of time invested by each team member in the process. Martech buying committees vary in size but average between 2 and 5 members, based on the project’s scope and the organization’s size. Most of the resources here go on vetting vendors and integrating them into the existing tech ecosystem.
Before deciding on either option, you must consider the resources available to you.
2. Experience – A second essential aspect is the experience in the organization. Has it built solutions outside its core product before, and was that endeavor successful or not? For example, at Lyft, previous automation investments generally resulted in a 20-30% improvement to the cost efficiency of automating activities. Building marketing automation solutions isn’t easy, from bringing together the customer data into a useful infrastructure to creating the needed tools for orchestration and attribution, the skills required are varied.
3. Time-to-market – Directly related to the two items above, time-to-market is a crucial factor. Although implementing sophisticated marketing automation solutions can take 6 to 8 weeks, building them might take longer. For Lyft, it took 60-70 work months, eight people working for eight months in parallel, to get the first version rolled out. If your organization is under time constraints to roll out a solution, time-to-market might be the defining factor.
4. Use-cases – Finally, as with selecting a solution, when deciding to buy or build one, having your core use-cases clear is of paramount importance for the success of the project. When evaluating solutions to buy, make sure to ask questions, and receive proof of coverage for your use-cases. While when building in-house, make sure the capabilities exist to provide satisfactory answers to the use-cases at hand.
When Lyft decided to build instead of buy, they focused on creating an optimized bidding machine that would adapt bids based on LTV calculations and the specific uncertainties and needs of each market they worked in. Only when they had the problem framed as a use-case did they understand whether they could build it and exactly how to go about it.
If you decide to build
Deciding to build a marketing technology in-house is a daunting task. Here are three tips to help make your experience smoother.
1. Work cross-functional – Yes, your R&D team is building the product, but they shouldn’t go at it alone. They might have never built a marketing solution before. Similarly, your marketing is the one outlining the requirements but might have never been in a Product Management role.
Building a cross-functional team is how you become more agile and help each team to focus on its strengths. For example, marketing can create user stories while a product manager translates them into actionable features. Engineers can build the product while marketers QA it as users. The synergies resulting from such a team are priceless.
2. Focus on incremental releases – Rome wasn’t built in a day, and your in-house marketing technology probably won’t be either. Identify the different value drivers you are trying to build and decide which ones to develop first, thus allowing for quicker course-corrections, on your way to gaining value faster.
Lyft is an excellent example of this. The fact that they already had specific capabilities built in-house made it a lot simpler to execute. If you are starting from scratch, identify those smaller pieces, and build up to your final product.
3. Stay open-minded – As you begin to use your in-house solution, or plan for second phases, don’t lock yourself into an always-build mindset. Evaluate each arising pain-point and ideal solution as if it were the same primal “build or buy” question. As Lyft’s engineers said: “eventually, we will plug-in best of breed commercial components, such as data management, particularly as we start to scale our programmatic marketing efforts.”
If you decide to buy
With all the noise out there, buying marketing technology isn’t much easier. Hopefully, these three tips ease the pain.
1. Focus on your use-cases – It is easy to get caught up in vendor hype. Don’t let experienced salespeople sway you away. Come prepared with the list of use-cases you have collected. Since no vendor can meet all your requirements, determine ahead of time which use-cases are mission-critical and which can wait. Use that list to vet vendors as you research and meet them.
2. Work as a committee – You probably won’t find yourself discussing this dilemma if choosing the right solution is easy. For sophisticated marketing technologies you are better off deciding as a team.
If you are a decision-maker but not the user, make sure to involve them early in the process. Users should be allowed to provide feedback on the use-cases at hand and try the product before the purchase. Similarly, if there are adjacent teams that will benefit from the new solution, invite them to vendor discussions to qualm any doubts and ensure you hit the ground running.
3. Talk to experts – Finally, don’t decide without consulting with experts. Traditionally this results in two different conversations.
A – With your IT team. Make sure they also have a chance to evaluate the shortlisted solutions to evaluate ease of integration with your current ecosystem. The last thing you want to do is spend countless months fighting integration battles.
B – With references. Ask your shortlisted vendors for customer references you can speak to and make sure they are relevant to what you are trying to solve. Speaking with current users might be the single best way to know whether a solution is right for you.
How to decide when to buy or build
There is no secret formula or universal truth as to when to decide to build or buy, but when coming across this dilemma, make sure to follow these guidelines:
1. Use-cases first – Yes, I’ve said this before, but I can’t stress it enough. Use-cases are the lighthouse when adding new solutions to your company’s marketing ecosystem. If you can’t think of some very impactful ones, then the technology simply isn’t worth it.
2. Determine your time-to-market – The allure of building something in-house is great; the question is, will it be ready when you need it? For Lyft, the alpha version of their bidding system took nearly eight months to be released. Enterprise marketing solutions should take 2 to 3 months to be completely online, with some being ready within days.
If you work within a firm deadline, you might be better off buying than building. The opposite is also true. If you have the skills and the time, building a marketing solution in-house, tailor-made to your use-cases, might just be the better route to take.
3. Estimate your results – After planning your use-cases and determining when you need to go to market with this solution, identify what results such a solution would bring. One way of doing this is by identifying previous similar projects. For example, at Lyft, they used the cost-efficiency improvement achieved in a preciously developed coupon-spend allocation automation as a proxy to expected the return for the marketing automation solution.
4. Evaluate the cost – Cost-efficiency should prevail. Determining the costs of each option goes a long way towards achieving it. Your cost-to-buy is pretty straight forward, but don’t forget to include the internal costs of integration. Your cost-to-build should take into account the estimated hours of work from the cross-functional team as well as any additional employees that might be required to pitch in. Together with your predicted results, understanding which of the two options is the more economically savvy should be more straightforward.
To read the complete Q&A with Lyft’s engineers, click here.
Rony Vexelman is Optimove’s VP of Marketing. Rony leads Optimove’s marketing strategy across regions and industries. Previously, Rony was Optimove's Director of Product Marketing leading product releases, customer marketing efforts and analyst relations. Rony holds a BA in Business Administration and Sociology from Tel Aviv University and an MBA from UCLA Anderson School of Management.
