Whenever a recruiter reaches out to me about a job opening that I’ll decline, I encourage them to use me as a resource in the future if they need information about the nascent fields of analytics, big data, and quantitative marketing for games. As a result, I am asked — infrequently, but often enough to be something I think about — about the requirements that should go into job specifications. The question I am most asked is: How well should a marketer be able to code?
The answer to this question obviously depends on the seniority of the marketer; a VP of Marketing absolutely does not need to be able to output “Hello World”. But from the Manager level down, I feel strongly that marketers should be able to code. This is because marketing for gaming is fundamentally different than other forms of marketing – and different even from marketing for some consumer internet verticals.
Games generate massive volumes of data, which is a boon to marketing departments at gaming companies for two reasons. The first is that large volumes of data create the potential to extract value from robust, targeted monetization and re-engagement strategies. And the second is that large volumes of data are easier to extract value from using unsophisticated statistical techniques; the smaller the dataset, the more effort must be spent manipulating and normalizing it. But unsophisticated as they may be, these statistical techniques must be well understood by marketers in order to be used correctly. And given the sheer volumes of data created by instrumentation in games, the appropriate tools required to cultivate actionable insight from these data sets are programmatic.
I’m of the opinion that, at the very minimum, a marketer should be able to query the data she needs, manipulate it without requiring assistance from an engineer or DBA, and produce a conclusion from that data using practical statistical methods. In practice, this might look like a SQL query and some commands in R or a similar tool (Matlab, Stata, SPSS, etc.). This kind of “coding” is purely functional and can be learned on the job or through personal interest; in other words, it’s not a lot to ask of a candidate.
Note: the level of statistical knowledge expected of a marketer is a different story. Properly understanding and drawing a conclusion from an A/B test requires a deeper understanding of statistics than mere descriptive statistics can provide. But I have already covered this topic.
I read two things today which inspired this post. The first was a tweet by an executive at EA Digital which indicated that every marketing department in his group now contains at least one developer (I asked him to please not allow those developers to call themselves Growth Hackers). The second was a post on Numerate Choir about the world’s first “Marketing and Analytics Analyst” position. I believe these two posts highlight the perception shift the field of marketing, especially internet marketing, currently enjoys.