Choosing a 3rd party service without losing your shirt

Choosing a 3rd party service without losing your shirt

Segment enables you to collect customer data with one API and send it to hundreds of tools for analytics, marketing, and data warehousing.

As digital product engineers, there are a few over-riding principles that govern our work. As an agency, one of the most critical of these is “build once, use many times.” Using existing building blocks for standard features allows us to develop rapidly while leaving budget for the innovative features that will actually differentiate our clients’ product.

In many cases, these building blocks are part of our own arsenal of code, standardized for reuse. In other cases, we evaluate 3rd party services that allow us to introduce production-ready features without the expense of developing and maintaining them.

By using 3rd party services, not only are we able to reduce development costs and get to market faster, we’re also able to satisfy user (and thus, client) expectations. So when a client wants to eradicate spam and fraud just like Krrb, Airbnb, Twitter and Shopify, we can recommend the predictive service of SiftScience. Or introduce type-to-search functionality just like Netflix? Algolia will do the trick. How about passwordless sign-in just like Medium? Auth0 can do that along with social login for every network under the sun. The list goes on.

The question is, what is the best approach to vetting a 3rd party service? Here’s some advice from the AREA 17 trenches…

1. Respect your roadmap

If you run a digital product, you probably come across cool new services each day. As you fight to grow your business, these shiny new tools can be tempting. Just like any other feature, you must add it to your roadmap and prioritize it.

Consider the following:

  • What’s the effort? The effort may be low for engineers to integrate it but do your community, marketing and sales teams have the time to activate and promote it?
  • What’s the risk? Every feature added to your product has a risk, whether it be convoluting your critical user path to opening the business to unwanted consequences.
  • What’s the value? Is this good for the user or good for the business? Or hopefully both? In what way? Will it increase revenue, conversions or engagement?

2. Do your research

Most 3rd party services are competing for the same value proposition. Compare each service side-by-side, get product demos and don’t forget to search Google for reviews and complaints. Quora and Stack Overflow are also good options for asking peers for advice (most often, a discussion already exists).

Consider the following:

  • Will the service be around tomorrow? Check CrunchBase to review their team and funding. If comparing two competitors, the one with the latest round of funding is probably a better bet.
  • Do they have good documentation? Speed of integration is dependent on how good their API docs are. The frustration of spending your week debugging only to discover it was a typo in the documentation is not fun.
  • How good is their support? While Stripe is one of the best payment processors with decent documentation, they take two days to answer operational support requests, urgent or otherwise. They do not offer phone support, let alone a dedicated account contact.

3. Reduce dependencies

Lastly, after you’ve planned your roadmap and done your research, consider these final three points:

  • Performance: Although most integrations load asynchronously they do add performance expense, some more than others. For example, AddThis for social buttons loads jQuery so if you already use it, you may be loading it twice.
  • Extractability: The day will come where you’d like to remove or replace the functionality. Make sure you plan your exit strategy from the get-go, including data-portability. For example, Disqus makes it pretty difficult to port your comments.
  • Data Unification: If you’re tracking user events and conversions (as you should be) then I highly recommend using Segment. This will allow you to unify all your data across all reporting platforms and give developers the debugging tools that’ll save hours of frustration.

As a final note, I cannot stress enough the importance of the second point (do your research). Start-ups go out of business every day. When choosing a payment platform for Krrb Classifieds, we narrowed the options to Balanced Payments and Stripe. As a last minute exercise of due diligence, we researched them on CrunchBase. We learned that Balanced hadn’t received investment for a couple years whereas Stripe’s investment was healthy and recent. This prompted further research which ultimately led us to choose Stripe. A few months later, Balanced announced the closing of their service.