One side effect of working in a rapidly growing industry is the copious amount of job interviews, both giving and receiving. In some of these interviews, it is necessary to determine (or prove) whether some one actually knows how a technology works, or is just really good talking about it. With ad tech, this can be pretty tricky because it usually requires a fair amount of experience just to know the lingo.
Unfortunately, having the vocab and knowing the names doesn’t always translate to technical know how. So if you need a real pixel jockey, how can you figure out if a candidate really knows their ad tech?
Short setting up an ad server, the best method is working through several real-life scenarios, each emphasizing different aspects of the ‘technical foundation’. If constructed correctly, you will have a really good sense of whether a candidate can ‘hold their own’ and where they may need training.
Here are some interview use cases which I think are fairly useful:
–An advertiser wants to run 2 different campaigns for people that have visited their site. Campaign A are for people that have visited the site in the last 3 days. Campaign B is for people that visited 4 to 8 days ago. No one should see both campaigns at the same time. How do you implement this with standard ad serving pixels?
–You are a publisher and your ad server is not RTB enabled. Your largest advertiser wants to selectively buy an audience from you, but there’s no data integration – so you have to implement a pass-back arrangement. Describe the steps to set this up. Extra credit: how to you prevent the dreaded pass-back loop?
–How can an ad verification or visibility service reduce the reported performance of a campaign but not the actual performance of a campaign?
–You are an advertiser running an campaign using RTB bidding. You have a very strict CPM limit for the campaign as a whole. Would you be willing to bid over this CPM? If so, how do you figure out what your max bid can be?
–A long running campaign suddenly sees a rapid decrease in performance, with lower conversion rates and increasing CPA? Assuming no technical malfunctions, what might have caused this?
–An advertiser is running an A|B test using online conversion as the success metric. Line A targets an audience using cookie-based information, line B is the control running RON. Does cookie deletion bias this test? If so, which way?
–You are at an ad network with 1000s of sites. An advertiser wants to use their TPAS to measure site-by-site performance, but they only give you one set of creative tags. How do you solve this problem with macros?
–An advertiser wants to leverage their offline CRM database to target their loyal users using online display ads. What steps or services do they need to employ to make this happen?
–Explain a case where raising the f-cap can improve campaign performance.
–If there is an unusually large delivery discrepancy between a rich media ad server and a DSP system, what is the likely cause? How would you address this?
These scenarios should cover a fair amount of the core topics in today’s ad tech space. Obviously, there are many skills not covered – video, rich media, etc – but if the candidate knows the fundamentals, these are usually easily learned.
Teams that interview really well will spend a decent amount of time brain storming and sharing various use cases to ask candidates – even assigning them to different interviewers as part of the interview process. Not only does this improve candidate selection, teams have higher confidence in new team members on day 1.
I’d love to hear any other scenarios, areas that you think I’ve missed, etc. And if you aced these, drop me a line….I might know some folks that want to talk to you.