How to Model Your Way to a Successful Product Launch
When marketers do marketing mix modelling (MMM), or more generally econometrics, they use past results to develop forecast models and optimise what to do next. That data may include weekly promotions, detailed campaign data across all the media channels the brand has used in the past, and granular store-level performance. It typically goes back 2-3 years to account for seasonality and other macroeconomic factors. Reliable historical data is critical for marketing mix modelling to be of any use.
But when a brand launches a new product, where’s the historical data? There’s no prior sales to work from, and no track record of past decisions to analyse marketing performance for winners and losers. How do you identify the target audience? What are the most suitable channels to promote the product right off the bat?
It’s a daunting prospect. In the post-pandemic economy, slow and steady doesn’t win the race anymore. No brand can afford to tip-toe around its new product introductions. But launching a new product doesn’t have to be a leap into the unknown either.
Enter meta-modelling
What if, instead of flying blind, or biding our time for a couple of years to accumulate enough data, or simply relying on proxies, we could put together a composite marketing mix model, or meta-model, out of models that already exist? Say, from a handful of sister brands, or from what we might be able to gather from competitors. Could we use industry benchmarks to simulate the market dynamics we expect to encounter with our new product? If those other models and benchmarks all point to TV, or social, or print magazines as a preferred channel, then we’re not starting from scratch.
This is exactly what some top automotive manufacturers have been doing recently to jump-start their Electric Vehicle introductions.
To pave the way for their entry into the battery EV space, a client of ours picked three MMM models (from existing nameplates in their lineup), adjusted a few key dimensions based on industry benchmarks and their intended audience, market and product features, and they immediately hit the road with a bespoke meta-model. From day one, they were able to fine-tune creative messaging for upper funnel (web visits) and lower funnel activity (unit sales).
It wasn’t a moment too soon either. With Tesla doing more than just knocking at the door—so far in 2021, Tesla has sold 2.5X more EV units than its closest competitors—the European market for EVs (both battery and plug-in hybrid versions) is heating up. It grew by more than 130% in 2020 (while the total auto market shrunk by 20%) and now represents 16% of total car sales.
Use benchmarks to pave the way
The success of a meta-model depends on being able to tie that new product to existing products. If it’s too radically different from anything that came before it, we’ll be hard-pressed to find reliable benchmarks for it. But there aren’t many product introductions that are so radically different that they leave marketers completely in the dark.
Take a new film release. By the time the studio and distributor start to promote the film, its genre, director and lead actors are a given. They have their fan base and favorite marketing channels. For example, in a multi-year study we conducted for Twitter a few years ago, we found that adding Twitter into the mix could boost a TV campaign’s ROI by 11% for an action film, and by 24% for a comedy. When you add this type of insight to the film’s release date, its geographic distribution, competitive releases, and other macroeconomic variables, you can develop benchmarks for just about every configuration.
Thanks to benchmarks, a meta-model is like Avengers: Endgame, where the Avengers and Guardians of the Galaxy are all joining forces to explore multiple timelines, create a new reality and save the day.
Why the rush?
There was a time when brands could afford to launch new products at a more cautious pace—test the waters, take their time to incorporate feedback, iterate. There's still a need for careful market research, of course, but everything happens faster these days. It doesn’t matter if it’s a new breed of electric vehicles, a new film, a new credit card, or a new D2C channel for meal kits, consumers expect to be wowed right away, and if they're not, they have access to many alternatives at the push of a button.
This urgency didn't start with the pandemic, but the past 12 months have taken it to a new level. Necessity is the mother of all inventions, after all. Facing prolonged lockdowns, a new wave of online shoppers has finally broken through the barriers that slowed down adoption in the past, like payments, delivery, and trust concerns. That sudden influx of online shoppers is shifting the balance of power in retail.
How sudden? A recent study on consumer behavior by NatWest and Retail Economics in the UK estimates that online sales just grew by five years in that 12 month span. They found that in the past year, almost half (46%) of all consumers completed a new online purchase that they previously only ever purchased in-store.
There's no going back.
Moving fast doesn’t mean throwing caution to the wind
Considering the uncertainty of launching a new product with no historical data to back it up, most brands in the past chose to innovate in small increments: staying close to the products they launched before, the customers they already knew, and the market dynamics they were accustomed to.
But today's market realities are forcing brands to make faster decisions and take on a bit more risk. Thankfully, innovative approaches like meta-modelling make it possible for brands to wander off the beaten track and keep pace with new opportunities without throwing all caution to the wind.
Want to know more? Please contact us and we’ll help you embrace the future of marketing mix modelling.