7

What is it like working on ads teams?

Profile picture
Senior Software Engineer at Redfin2 years ago

I'm considering joining a new ads org, and I was wondering what working in ads was like overall since I've never worked in ads before.

A more specific question: There's a lot of competition in the online ads space nowadays - Will that make ads orgs more prone to layoffs as they find trouble gaining traction in this crowded market?

1.3K
5

Discussion

(5 comments)
  • 12
    Profile picture
    Robinhood, Meta, Course Hero, PayPal
    2 years ago

    I worked on ads at Meta, so I'll share my experience here. Some of this is specific to Meta, but I do think there's some common themes here:

    • There's more non-engineering complexity - This is due to there being an entirely new party: Businesses who buy the ads. So you still need to care about the end-user since they see the ad, and now you have this other, much more important party (the businesses giving you $$$ to run ads). This can lead to more roadblocks shipping projects as you need to make more parties happy.
    • Stress is higher - It's much worse having a bad experience for a paying enterprise customer vs. a free user who uses the app/website. This means that issues are treated with higher urgency. Back-end oncalls for ads are some of the most stressful oncalls out there (I worked on mobile luckily, so it wasn't too bad). Zooming out though, I worked for Meta so ads is way higher-priority (literally ~97% of company revenue), so if you're joining a newer ads org, I imagine it won't be as bad since ads won't be as big a portion of the overall company pie.
    • Figuring out impact is more straightforward - The success metric for ads is obvious: Revenue. When we were scoping out projects on my ads team and then talking about them in performance review, it was pretty easy to see where you stood as you could do math to figure out how much new revenue you brought in. I feel like this makes paths to senior/staff/beyond clearer as the success criteria is so obvious: Bring in more $$$.

    In terms of layoffs, I understand the point about competition, but I feel like this is true for a lot of orgs as tech companies are copying each other all the time. I feel like the average ads org is more insulated from lay-offs as they directly bring in revenue. Any sizable tech company will have tons of teams working on features without as much traction which don't bring in revenue directly - Those orgs will probably be cut first before ads.

  • 9
    Profile picture
    Robinhood, Meta, Course Hero, PayPal
    2 years ago

    Good questions, but they're tricky as every ads team will be different. If I were to pick something, it would be to always have the advertiser at the front of your mind. The advertiser is the most important user in the ads equation, so you want to make sure you have the infra necessary to keep them happy. Here's some examples:

    1. Great analytics so they can see how they're doing
    2. An ultra-fast and smooth pipeline for them to report bugs
    3. Clean tooling to allow them to "preview" how the ad will appear to the end-user before they launch their campaign

    The bar for all the good software traits (robustness, reliability, scale, etc) is much higher with enterprise software, which is what ads effectively are. With consumer software (e.g. the features for "regular" everyday users in Meta apps), you can cut corners and ship a lot of things fast with rough edges - You can't really do that with ads tech. Make sure you're always fighting for quality when doing ads product launches; this is very much inline with what it takes to grow to a senior+ engineer anyways.

    For folks interested in leveling up their code quality game, check these out:

  • 4
    Profile picture
    Senior Software Engineer [OP]
    Redfin
    2 years ago

    How can you help your team to build an exemplary architecture for ad teams? Or what are the design challenges and how can you step in as a senior engineer and solve these complex issues especially when your team is new?

  • 4
    Profile picture
    Friendly Tarodactyl
    Taro Community
    8 months ago

    Two years after the initial question - wondering how OP has been so far in your team?

    For me, I've been in advertising from a big tech for two years. Advertising business is huge and complicated. As with any big tech products, there's all kinds of teams (all kinds of offline pipelines moving data around, caches, analytics, runtime bidding systems, signal ingestions, etc...). One thing I like is that the domain knowledge seems transferrable at least on the surface level. I interviewed with a few other ads teams from other companies and they seemed to enjoy our conversation and we can cut many context setting sentences, because they know the common terms.

    One thing I find advertising different from my expectation before I join is the so-called crazy low-latency runtime bidding system. I feel my previous mobile team has better engineering insights, knowledge and optimizations and my current org is just business logic focused... Thinking of the broader picture, many other non-ads systems have low-latency requirement as well, e.g. google search, amazon shopping checkout, fluidity of any mobile apps / webpages

    Back to the layoff questions, I think all you need is just a crystal ball and no one really knows. Because advertising is huge, laying off certainly could happen. I do agree that advertising is so much more reluctant to organization shutdown as compared to other orgs, because it's usually making money and if it's not, executives have more patience to invest and make it profitable.

    • 2
      Profile picture
      Tech Lead @ Robinhood, Meta, Course Hero
      8 months ago

      I don't think OP is active on Taro anymore, but IIRC, they got laid off from that ads org 😢

      I agree that a lot of ads context and skills are transferrable. For me, the huge emphasis on data analysis and logging was huge. Instagram Ads was the first engineering org that pushed me to get really good at SQL (I've written 100+ line queries parsing through trillions of rows), and now I'm always thinking about how I can actually tell if the end-product is delivering user-value.