The algorithm has gotten a lot better at forecasting.įor example, when COVID-19 hit, a lot of people started cutting back on their budgets as a knee-jerk reaction. We’ve been doing this for almost three years, which has allowed us to build up a lot of historical data. Has this approach helped you during the pandemic? We’re constantly rotating images and calls to action to try and figure out the right personalized message for any given user on Facebook, for example, which is a little different from Google or Apple Search or any other partner. We do a heck of a lot of creative testing. We ended up hiring more people in house to support our creative inputs. We might have spent $700,000 or $800,000 with Facebook, Google or any given partner. before is that we used to have fixed budgets. Once we were able to automate that, we could run tens of thousands of A/B tests over the course of a month. Before, we were able to run a couple of hundred A/B tests with the team I had. We’ve been able to increase the velocity of our A/B testing and our rate of learning to find the right audience, messages and experiences that bring people into the product and create the habits that lead to the best user lifetime value (LTV). We set up API integrations and worked with them to start automating all of the levers that we could control, including bids, budgets and orchestrations. I found a company called Nectar9, which provides AI-powered marketing automation technology. I wanted to have a machine do it and keep our team lean from the start. I wanted to build something for us that would let me see how my data is doing across the ecosystem without giving away control.Ĭompanies like Uber and Lyft have developed similar capabilities, but they also have huge teams to maintain them.
It was all about you spending more with them, and we had little purview into how our Facebook campaigns were doing compared with Google or other platforms. I got into a lot of the Google, Facebook, Snap and even TikTok betas as they were developing their automation for things like Universal App Campaigns, and as a beta partner I learned a lot about how they were building their machine learning capabilities to optimize the data they were getting from us.īut I came to realize that a lot of the benefit they were seeing came from a self-serving algorithm. It’s just a matter of having the AI to automate the triggers for when to buy and when to sell. In my case, the return is return on ad spend or the cost to acquire a customer. We both spend a budget and try to invest with a return. LOMIT PATEL: I began studying the finance industry and the big brokerage houses and realized that the way they do what they do is similar to what I do in UA. “It was a good time to spend and get users in,” Patel said, “and a good time to look at increasing our footprint outside of the US.”ĪdExchanger caught up with Patel to talk about how IMVU is using automation to keep growth humming during COVID-19.ĪdExchanger: How did you get started with automating your UA? Which is why IMVU pulled a large portion of the budget it had earmarked for the second half of the year into Q2 when pricing was still low. “We’re going to see special interest groups flood the social networks as a way to influence voters, and that will lead to a natural increase in costs on Google and Facebook.” “In most cases, marketing budgets are use it or lose it, so I see a big flood coming in Q4, not to mention the upcoming US presidential election,” he said.
More recently, when COVID-19 hit and advertisers started pausing their ad budgets, Patel started thinking about Q4. Three years ago, that meant introducing a high level of marketing automation into IMVU’s user acquisition (UA) efforts and focusing more on IMVU’s mobile app than its desktop presence. “Whether things are going well or not, you have to focus on predictability in driving customers and revenue.” “I never only think about the moment, I think 12 to 18 months out,” said Patel, who joined IMVU in late 2016. Growth marketing, he said, is as much about expansion as it is about mitigating risk.
Lomit Patel, VP of growth at IMVU, a 3D avatar-based social networking universe, was already planning for an economic downturn in 2017.
0 kommentar(er)
