Right before Halloween, we asked UA managers about their worst work-related nightmares. Read on to learn about some of the scariest fails and mistakes you really want to avoid making yourself.
#1 Emptied Bank Account
"I have a UA Nightmare from Fumb. Using Facebook’s CPCV to run a campaign and spending $1000 in 5 minutes. As a bootstrapped studio, this was not great.
I set it up as I used to work at AdColony where we ran CPCV (Cost Per Completed View) before CPI was even a thing. At AdColony a CPCV was that, a completed video of 30 seconds. At Facebook, a CPCV was 3 seconds of a video. So it just spammed out and spent loads super fast. We got zero installs. To be fair to Facebook, they refunded it. Thank goodness as it emptied our bank account at the time."
#2 Automate & Forget
"This was during the early days of Google introducing UAC back in 2015 - times were changing for app install campaigns. We had a fairly niche app that was struggling to perform with UAC - spend was definitely there, but the return was an issue.
As an impressionable Junior UA Specialist, our AM's word was gospel. When they recommended increasing budgets 'til it worked, I jumped at the opportunity to spend. Being the tech-savvy up-and-comer I was - I decided to do a big brain move and set up automated budget increase rules. A 15% a day increase when ROAS was under target sounded like a very good idea.
We ended up attending a Google event - most likely App Summit, for the rest of the week in Dublin. Now, this was incredible! Visiting the Google office was mind-blowing. So mind-blowing that I completely forgot I had this automated rule set up. You can probably see where this is going.
The rule ran every day for a week straight and we ended up spending A LOT. I had a lot of explaining to do, but it was all taken pretty well - we approached it as an ad hoc experiment and ended up getting a lot of useful learnings.
Main takeaway; if you're going to use an automated rule, don't forget about it.. and if you do, be prepared to give it a positive spin."
#3 No SDK
"One of the biggest nightmares I lived through was related to an MMP error. We were in the process of switching MMPs and apparently we once A/B tested a version of the game without an SDK on 5% of the users. Such a low percentage went unnoticed. But as the games’ versions kept updating, 5% became 10%, then 20%, then 30%, and so on. As this game was on a declining trend, we did not notice this tracking issue and could not understand what was going on with UA. I spent a month raising bids and refreshing all the creatives as often as possible to get the spend back up. Of course, it didn’t change anything: as some users had no MMP, that meant networks didn’t receive postback triggers, and could not increase the spend.
Finally, with the last version update, 100% of users weren’t tracked and all the campaigns on all networks stopped. When we asked the developer if he could double-check if we had any issue with the MMP SDK integration, he answered directly: 'Oh, but we don't have it in the game'.
It was the worst, it almost made me cry to see that I had worked so hard on this while it simply couldn't have worked since we didn't have the MMP.
Lesson to be learned: work as close as you can to the product team."
"We had an unfortunate experience. I will not mention the partner, but we were working well with them, the costs per trial were very good and perfectly targeted in the geo we wanted. But we always have to look further down the funnel. After two weeks of data, we realized that the conversion rate after the trial period was 10 times lower than the average. Fraud? We'll never know. But be careful, looking at just one KPI doesn't work."
#5 Dot or Comma?
"So, I was working until very late hours at an ad network, and I had to start the WW scale for one of my Hyper-casual clients before the end of the day. As you can imagine, the CPI bids were pretty low for the WW targeting and since it was a scaling phase, the daily cap of the WW targeting ad group was around 10k. Our system was updating bids from an Excel spreadsheet upload. So, before going to bed, I uploaded the bids taken from my client and everything was going super smoothly. The next day, I woke up with a message from our VP of Growth who lives in the US. He was asking for the reason for the huge spike in this client’s spend. Our algorithm was pretty slow, so I wasn’t really expecting such an impact. I went to check the campaign straight away and noticed my mistake when uploading the bids with the excel sheet. Instead of using DOTS (.) I used COMMAS (,). So basically all the bids were multiplied by 10!
I refunded the client and had some explaining to do."
#6 Test Ads
"I think one nightmare that still haunts me from time to time is iOS14.5 thanks to the abundant amount of Ad level and user-level data reported back to advertisers that's also delayed up to 72 hours and often incomplete under SKAN framework. It has definitely made testing and analyzing the performance of creatives a bit tricky. For some time I continued to upload and test new creatives on iOS and Android just like pre-iOS 14.5 but at some point, it’s become hard for me to understand which creative is actually driving the performance on IOS and which creatives are just eating up the impressions. Instead, I have now switched to testing new creatives first on Android and using the IPM, CTR, and CVR from Android as the benchmark."
Are you scared yet? Here are a few more from last years' nightmares...
#7 $17K Overnight
Kevan O'Brien, Manager, Ad Monetization and Optimization at Daily Hive
"On my first week at a studio I used to work at, I noticed that one of the ad sets in one of my campaigns was exceeding goals, with a 100% return in just a couple of days. I got so excited that weekend that I changed the budget from $50 to $100 to double down on the success. It wasn't until Monday morning when one of the engineers tapped n my shoulder and asked, 'Did you know we spent $17,000 on one ad set overnight?' I quickly checked and realized that the 100.00 cap punched in as 10,000. Probably something to do with decimal points. The campaign ran for an entire day and a few more hours the next day before we caught it, spending $17,000. I wasn't used to working with such high budgets before, and I was sure I would lose my job. We had an emergency meeting with the CEO and Marketing Director, where I explained how I made this mistake. Then my CEO at the time said, 'Ok, let's track it. Let's learn from it.'
We got 5000+ installs from our top-performing ad set, and I expected to get good quality installs from that cohort. And they were. It took us a bit longer than seven days to hit our target, but looking back, after three months, we tripled our revenues.
We learned something from this that not many studios had a chance to learn: what happens if you 1000X your budget overnight?!"
#8 Devaluing Game's Economy
Farhan Haq, Head of Growth at SYBO
"From the perspective of User Acquisition, Monetization, and Publishing, my worst nightmare would probably be the product team devaluing a game's economy in an otherwise successful soft launch with strong retention rates. The lesson: don't go too crazy with changes even in soft launch. If metrics look solid, just launch the game!
Generally speaking, and to be more practical, I would advise not making more than two changes at once on the monetization of a game."
#9 Postback Attack
Stephanie Unger, Director of Advertising Operations at Blind Ferret
We run a rewarded engagement platform, where users can collect redeemable points for completing various events in our client’s apps. Since we are only paid for the initial install, the number of points awarded for each event is determined by how frequently we expect a user to complete it - obviously, it would cost us a lot of money to hand out too many points for a goal that almost every user achieves.
We launched a campaign while this model was still very new to us, and we were under the impression that we were rewarding an extremely long-tailed event with a very low conversion rate. As we normally would, we offered tens of thousands of points to any user who reached this level. Little did I know, the client implemented the wrong postback and was firing an “app-open" event instead - virtually everyone reached this goal and the rewards were handed out like candy. We only noticed when a very honest user informed us that she received her reward much earlier than anticipated. Who knows how long this would have flown under the radar if it wasn’t for her. I remember losing sleep over this mistake - how could I not have double-checked before launching? How was I going to tell my boss? At the end of the day, it was a very valuable lesson. We now ALWAYS double-check for this sort of thing, and have a system in place to notify us when rewards are costing us more than we are making."
#10 Wrong Country
John Wright, VP Operations at Luna Labs
"My worst UA nightmare is probably related to not double-checking bid changes. I've had situations where I analyzed the data and made a CPI change based on evident results, such as higher than average ARPU, for example. But instead of updating the correct country, I've changed the bid of the wrong one.
It's quite easy to confuse IN (i.e., India) with IE (i.e., Ireland); The result is a bid that is five times higher than average for the market. Subsequently, you either burn out your budget because you're bidding too aggressively, or end up making a significant loss on all the users you acquired because the ARPU of that market is far less than the one you intended, or both. I think every UA person has done this at least once in their career. So always double-check your bid changes!"
#11 False Promising
Sun Baek, Performance Marketing at Trade Republic
"A while ago, I started a partnership that did not work out because the partner overpromised on things they didn't have yet. We were looking for traffic in Germany, and this company presented us with numbers of audience reach and daily/monthly traffic and promised to deliver a certain amount of subscriptions. So we signed, waited, and... nothing happened!
This experience taught me to do more background research before starting a new partnership, always expect the best but prepare for the worst, and ask the question, 'what happens if the goals are not reached?'. It also taught me to start with a smaller test budget before fully engaging a new channel and to evaluate results before scaling a campaign."
#12 An Extra Zero
Lorenzo Rossi, Co-Founder & Head of Growth at REPLUG
"I remember nightmares from the times I worked as a hands-on UA manager, responsible for six figures' monthly budgets across different advertising channels. I was always scared to set the wrong bid or budget on my Facebook or Twitter campaigns: adding an extra zero at the end, pressing Enter too fast, things like that. At that scale, the risk of burning marketing dollars due to human error was tremendous... I learned never to launch a campaign in the evening before leaving the office or on Friday just before the weekend, always double and triple check the setup, and keep an eye in case something looks wrong.
Another nightmare of mine is relying on false data. It happened that I had taken a decision based on data that was completely wrong, which in some cases resulted in massive spending for no good reason. So although we should aim to have a single source of data that we trust, it is crucial to access additional data sources to compare numbers and flag any technical problems."
#13 Quick Fix
Hannah Parvaz, Head of Marketing at Uptime
"A few years ago, I was brought in to help a company who had started to struggle with their Facebook Ads. Over the last month, they'd seen their costs shoot up dramatically (over 1000%) and couldn't work out why. I went about with a creative audit, a bunch of digging through conversions, and then looked into the targeting to notice that they were targeting previous conversions instead of excluding them! It was the quickest fix I've ever seen."
I hope you enjoyed the stories! Wish you a very spooky Halloween and hope that UA will never give YOU nightmares!