Read the full transcript
We are witnessing now a very interesting phenomenon. It used to be that the marketing team would own the decision on marketing products and what solutions they would use.
But engineering and product teams are becoming more and more dominant within marketing organizations, because first of all, marketing teams they understand now that they rely a lot on product and engineering to create unique experiences for them.
The second part is that everything becomes much more quantifiable and digital, so you want to be able to attribute all the different experiences and to see which works, which doesn’t work. And everything became just more, kind of, techy, more geeky and this where the engineers really, they become much more powerful.
It leads to an interesting phenomenon where now a lot of our smart costumers, they have the build versus buy decision.
So, there’s a tendency for engineers, and I’m speaking as an engineer, we want to build. The problem is, how do you maintain this? And will you build the features that your marketing team wants but they are not sexy and they are not interesting?
So, a lot of our customers, when they have a consideration on build versus buy, when they go deep into our APIs and our SDKs, the engineering team realizes that we can just become a building block in the experience they want to build and we literally are saving them two years.
We have over 60 engineers working on creating progressive technologies so every year we have 60 man years devoted to advancing our products.
Not a lot of companies can devote 60 developers for this type of solutions.
Led by baby-faced sharpshooting, the 2014-15 Golden State Warriors were America’s sweethearts. The nation loved that the team’s core group of stars (Steph Curry, Klay Thompson, Draymond Green, Harrison Barnes) were homegrown talent, a nucleus assembled over several years. In an era defined by buyers, the Warriors were consummate builders.
The same spirit that champions building product in-house transformed that valley south of Oakland from quiet suburb to birthplace of modern innovation. It’s ingrained an entrepreneurial ethos in industries from technology to eCommerce and stocked companies with engineers who almost always prefer to build.
Personalization has become so strategic for many marketers that an overwhelming impulse to build the technology in-house is sweeping over the market. However, building scalable personalization is a daunting task that requires vast engineering resources and years of expertise to successfully execute.
So, should marketers build or buy the technology to personalize experiences for your customers? There’s no simple answer. It depends on the size and maturity of the organization, skillset of a company’s engineering team and ultimately the end goals a marketer hopes to accomplish by deploying omnichannel personalization at scale.
However, years of working with some of the best engineering and marketing organizations around the world suggest that the savviest companies tend to do a bit of both. In deciding whether to build personalization from scratch or partner with a vendor, there are six main factors to consider. While this buy vs build decision framework is geared towards personalization, we hope it will prove helpful as you ponder the build vs buy decision across your organization.
- Time to Market– Will purchasing or building in-house allow me to deploy personalization faster?
- Total Cost– When all is said and done, is it cheaper to build or buy?
- Maintenance of the Technology– If I build in-house, how I will I maintain a best in class product?
- Workflow– Is it easier to manage a vendor or team that is building personalization tech?
- Scalability, Customizability & Performance– Will building technology in house provide better long term results?
- Expertise- If I choose to buy, what other factors should I consider besides strength of the technology?
Let’s examine each of these in a bit more detail.
Build vs Buy Software – Pros and Cons
Time to Market
Given that personalization has proven to deliver 5-15% uplifts in revenue, every day that the technology is not ready is money lost. Hence, speed of deployment is essential.
Many companies fear purchasing technology due to complex integrations that take months before the end user sees any benefit. While this can be true when trying to patch together point solutions, unified platforms can be fully deployed in weeks or even days. At Dynamic Yield, we had live use cases up and running for a major apparel client in time for Black Friday, even though the partnership wasn’t formalized until early November.
Even the world’s largest retail engineering teams can rarely match that speed. Thus, building from scratch in-house requires extensive planning and often working with a vendor at least in the short term to ensure your onsite experience isn’t just powered by a CMS.
While pricing specifics are quite intricate, most personalization technologies operate on some type of SaaS pricing model where the cost of the product is dependent on the volume of traffic to the website and quantity of services offered. For SMBs and mid-market companies, buying personalization often equates to less than hiring even one quality engineer.
However, enterprise deals for personalization are quite large, easily climbing into the hundreds of thousands of dollars. For teams with more resources staring down a large commitment, the temptation to buy grows. But consider this- the absolute minimum viable team to build enterprise-grade personalization may consist of one senior engineer and five junior folks and salaries of $200K and $100K each respectively. Working efficiently, they might be able to bring a platform to market in six months. Total cost in labor alone has already run you $350K just for an MVP!
Of course, the process of sourcing and evaluating vendors also carries a non-trivial utility cost. But for the price of devoting an internal engineering team, you can often purchase a platform with years of intellectual property and engineering power behind it, then put your team to work maximizing the technology.
Personalization, Recommendations, Behavioral Messaging, Testing & Optimization in a Single Platform
Building a personalization stack is merely the first step in delivering superior experiences to your customers. The technology needs to be continuously maintained while new features are shipped to keep pace with the competition.
It’s time-consuming, cumbersome and expensive. While shipping new features is exciting, maintenance of existing back-end technology can be a tough sell. Will your most talented engineers really want to work on patching bugs in recommendation widgets rather than building your newest retail VR experience?
In the video above, Dynamic Yield CEO Liad Agmon notes that companies like Dynamic Yield have “over 60 engineers working on creating personalization technologies. So every year, we have 60 man-years devoted to advancing our products. Not a lot of companies can devote 60 developers for these kinds of solutions.”
When considering whether building or buying a personalization solution will create smoother work streams, it’s important to examine how the stakeholders in marketing technology have changed. Marketing teams uses to have full autonomy over buying tech but increasingly CIO’s are becoming the ultimate enterprise decision makers. According to Agmon, “Marketing teams understand that they rely on product and engineering to create unique experiences for them” As everything becomes more quantifiable and digital, things get more techy, more geeky and engineers become more powerful.”
This makes sense- crafting better customer experiences is a cross-functional responsibility. Personalization technology has a direct impact on non-marketing functions such as merchandisers, product managers and analysts as it can automate some of their ongoing tasks and help make better data-backed decisions.
In the context of build vs. buy, this adds complexity whether you are purchasing or constructing technology. Buying will involve many stakeholders and potentially competing opinions on which product to buy. However, once a decision is made, a reasonably small group of end users can settle in and learn the vendor’s technology. With an in-house solution, several stakeholders will often have to devote intellectual horsepower to each new feature that is created. This can result in great technology that caters to needs across the organization but requires very deep commitment from busy people.
Scalability, Customizability & Performance
While there may be logistical advantages to partnering with a vendor, nobody knows your business better than you. For a technology as intimately intertwined with your product as personalization, there is no way an outside company could understand your customers well enough to build technology that performs as well!
Turns out this is less straightforward than it seems. It’s important to remember that build vs. buy is not a strict binary. The world’s best personalization technologies stand out precisely because they are customizable to the needs of each specific business.
Since many early personalization vendors were black boxes, a misconception formed in the market that buying meant losing control of crafting your customer experience. This is no longer the case. While building a software will always provide the most customizability, the next generation of personalization operating systems is all about open architecture, combining the scale of having 50+ engineers building new products with the flexibility for customers to easily customize their platform.
The Expert Quotient
In rapidly emerging or new areas of business, the advantage of learning from the best can never be overestimated. Systems like CRMs and Marketing Automation platforms became successful because companies soon realized that irrespective of what unique problem they thought they faced, somebody had already faced the same problem – and had devised a solution. This resulted in the growth of B2B communities, peer to peer knowledge sharing and an abundance of available talent that specialized in these tools. As personalization tech matures, we are reaching a similar stage in the business cycle.
At Dynamic Yield, we have the advantage of working with hundreds of brands across the world, giving us expertise that only comes with years of trial and error. How do you personalize experiences for low bandwidth environments? We have customers in Asia who have solved that problem. How do you personalize experiences when you serve customers across climates, affecting the products you have to merchandise? Our Russian customers have solved that.
Thus, partnering with a vendor isn’t just a matter of buying technology. Choosing a personalization stack with years of proven success in the market is also an investment in expertise that will pay dividends, even for features you choose to build in-house.
Conclusion: Buying Your Building Blocks
Long heralded as the golden standard of building personalization in-house, Amazon has spent years refining its technology, using a small army to do so. At the time of publishing, they have 36 open jobs for engineers on their personalization and recommendations team alone! However, it is possible for enterprises with smaller engineering teams to compete.
By purchasing an open-architecture solution, you can save your team years of development time, freeing them from building basic infrastructure. The extra development time can be used to adapt the solution to your particular needs and keep your engineering team focused on interesting and value-additive work while the vendor builds powerful machine learning algorithms.
As personalization becomes a C-level priority, there is a natural tendency to feel uncomfortable placing something so essential in the hands of a third party. But remember, when you bring on a personalization vendor, you aren’t outsourcing personalization or data science – just some of the infrastructure behind it.
“When customers go deep into our SDKs and APIs, they realize we can just become a building block in the experiences they want to build and we literally are saving them two years.”
Personalization, Recommendations, Behavioral Messaging, Testing & Optimization in a Single Platform