Build vs. Buy: A First-Hand Analysis

The social/mobile gaming space is relatively new, yet the evolution of the space has been rapid. Despite ever-changing tech and design trends, using analytics as a tool to drive success in this space remains critical.

Newcomers to social/mobile gaming must become overnight experts or risk becoming overnight failures. The first wave of successful social gaming companies, such as Zynga, Playdom and other innovators invested heavily in their analytics and business intelligence (BI) tools from day one. Over time they have built large analytics/BI teams, scalable infrastructure and robust A/B testing tools; all of which give them the ability to quickly optimize core game metrics. These companies, now behemoths in the industry, made massive investments in analytics from the start because no commercial solution specific to “social” existed at the time. These sophisticated tools, built from the ground up, gave them incredible insight, and a massive advantage over second- and third-wave gaming companies.

With new platforms being developed every day, and previously untapped “gamers” entering the space through free-to-play channels, the opportunity to build successful games is bigger than ever. So, how do up-and-coming companies get a comparable level of analytics into their games? Do they build it themselves or buy something commercially available?

I have worked in social gaming in a first-wave company (where we built our own analytics) and in the second-wave where we decided to buy an off-the-shelf solution. The goal of this post is to outline some of the factors to consider while making a Buy vs. Build decision at this point in the industry’s life cycle.

Social gaming metrics are now standardized:
Just as we have standardized metrics for website analytics via Omniture or Google Analytics, social gaming now has a set of metrics that are becoming best practice in the industry. (To learn more, download our whitepaper, The Top 7 Metrics of Highly Successful Social Companies.)

We know that 1-, 3- and 7-day retention matters. We know virality should be tracked on a per-viral channel basis. We’ve realized that every single step of a player’s first-time experience needs to be tracked and scrutinized. We understand the importance of a robust events system to track arbitrary actions in the game. Commercially available tools have evolved to provide these basic metrics and technologies. Imagine the power of having parity in analytics (with scale) with your competitors with rudimentary instrumentation. Two to three days of development effort gets you 80 percent of the way—that is huge.

Focus on the quality of the games:
The competition is stiff, not only from the established players like Zynga, but from smart, scrappy start-ups that are willing to launch fast and iterate. Facebook has simplified access to users and reduced barriers to entry. A sure sign of social gaming’s maturity; the best and brightest from the traditional gaming industry are entering the space.

It’s no longer enough to build a sub-standard game and drive traffic to it. Player expectations are rapidly maturing too, and to survive companies need to deliver on their demands. Building robust analytics capabilities that include segmentation, funnels, cohort analysis, while scaling and meeting up-time commitments is a daunting project that requires a lot of planning and resources. Because robust tools are commercially available, companies should put careful thought into whether it’s worth undertaking a massive (and ongoing) infrastructure project like analytics.

Data should be accessible across the organization:
Off-the-shelf tools provide robust custom dashboards that fit the entire organization’s needs, from the C-level to the Marketer to the Developer. With minimal training, each stakeholder can quickly create the view they need from a universal data set. A custom solution, much like any other product requires trade-offs, such as:

• Which views will your engineering teams focus on first?
• What part of the org gets priority?

When you work with a SaaS solution like Kontagent, the most important views are ready to go off-the-shelf and the custom views can be created quickly without engineering resources.

Consider resource and opportunity costs:
The cost to build your own system depends on the features that you think are necessary. Generally, companies start with building an event tracking system with a  third-party graphing tool as the front end. This by itself is not a massive undertaking and can be accomplished within two to three months (including instrumentation in game), with probably two highly-skilled engineers. However, building a full-featured system as offered by Kontagent—which includes funnel analysis, cohort analysis, ad-hoc math on custom events—is a large engineering project that needs strategy and thorough planning. That could easily take several quarters with a dedicated team of engineers, product managers and testers.

Additional reports, custom views, scalability (based on increased user bases) can increase the total cost of ownership of a custom analytics solution. At the minimum, you will incur engineering, DBA, operations and hardware/software costs associated with additional reports, maintenance and enhancements to your analytics system. The actual cost varies depending on your analytics needs, user base and the skill sets within your organization; the reality is that there is a significant cost (resources and opportunity). Third-party analytics tools eliminate these undefined costs for you and provide strong service level agreements (SLAs) that guarantee that you’re reporting and analytics system is working 100 percent of the time.

Domain expertise gives you a competitive edge:
An additional benefit of working with third-party analytics companies is access to data scientists who can help you make sense of all the data that your game is generating. Whether you’re a new studio creating your first game or a larger publisher building your first social or mobile game, this service can be invaluable in sifting through the noise and focusing only on the data that is relevant to the games current place in its life cycle.

Sometimes a hybrid system makes the most sense:
There are always cases where companies need a hybrid system: one specifically created for the business intelligence team and one for the rest of the organization. The decision to use one or both solutions depends on the complexity of the interactions in the game and specific correlations that are being sought. It also depends on the analytics requirements across your organization, and the level of expertise you have dealing with mining raw data.

In summary, third-party analytics packages have matured over the last couple of years so that the benefits—the ease of implementation, their ability to scale and their feature sets—far outweigh the cost. The focus in social gaming has shifted from a drive to build the best analytics tools, to a drive to build the best games. At the end of the day, it will be up to the individual company to determine whether they are capable of doing an excellent job at both, or if they are better served focusing their efforts exclusively on producing the next-generation of hit social/mobile games. As it stands now, if new companies want to get a head-start, they would be well advised to trust a company like Kontagent to provide their analytics to not only get off the ground, but also to continue to be successful.

——————————————————————————————————————–

About our guest author: Rohan Sardesai has held various product management roles specializing in advertising, mobile, social apps and games. He is currently consulting with multiple companies in the music, mobile and social apps space. Most recently he was senior director of product management at iWin, Inc., a leading developer, publisher and distributor of social and mobile games. Prior to that he was director of product management at Lolapps, Inc. If you have any questions for Rohan, feel free to email him.