It’s easy to hit the pause button on a performance marketing campaign. Maybe a little too easy.
Everybody wants to see return on ad spend (ROAS), and when it doesn’t happen right away, alarm bells can go off.
Those alarm bells take the form of early numbers you see at the start of your campaign. And we’ve all been trained to trust the numbers. Data-driven decision-making has become a mantra – not just in advertising but in every other area of business.
Relying on gut instinct is obviously not the answer. The problem is that no one really quantifies how much data is enough to drive the right decisions. As a result, there’s a real risk of playing it safe by stopping a campaign before it has a reasonable chance of success.
It’s kind of like driving down a long country road and assuming you’re lost, or at least nowhere near your destination.
Then, when you drive just a little bit farther, a few road signs appear.
Some landmarks you expected to pass by emerge over the hill.
Suddenly, you realize you’re getting where you need to be.
While cars come equipped with GPS to avoid this problem, performance marketing requires training yourself to look for the right signs – and not panicking when they haven’t shown up quite yet. Here’s how:
1. Avoid the ‘Day 1’ blind spot
Most companies took years to build their brand, through a series of campaigns and strategic actions that earned them consumer trust. You don’t get there on Day 1 of opening a business.
The speed at which we can now automate advertising, however, sometimes creates an expectation that results will come almost immediately.
Let’s say you launch a performance marketing campaign, for example. On Day 1, you see that your ad drove 100 conversions. You’re at 80% of your ROAS goal – hypothetically.
The truth is that 100 conversions is a small sample. Day 1 performance, while important, is not always indicative of future success. Changing a campaign prematurely based on initial results might mean leaving money on the table. You’ll make better decisions when you’ve reached 1,000 conversions, 10,000 conversions and beyond.
More data on your customers’ preferences provides a clearer and more consistent picture of market opportunities. Making the decision to pause your campaign or lower bids based on Day 1 results is like projecting the flavors of an entire ice cream business on the preferences of just the first three customers.
2. Guard against availability bias
The temptation is to jump on whatever data is on hand, a fallacy known as availability bias. It’s like seeing a broken window in your neighborhood and assuming the crime rate is skyrocketing. Law enforcement, however, usually spend a lot of time gathering more information before taking any drastic measures.
Most businesses are similarly cautious with customer feedback. If someone calls into a contact center explaining they couldn’t get a store associate to help them find what they were looking for, the agent would probably apologize. Only if dozens or more customers called in with the same complaint would a retailer begin hiring more staff, investing in more training or creating more visible store signage.
You can avoid availability bias in performance marketing by taking your desired campaign objectives and determining what constitutes a meaningful frequency of conversions. These could include outcomes like:
- E-mail signups
- Form submissions
- Purchases
Ask yourself what volume of conversion data would justify making changes. In other words, when would you feel confident in optimizing, iterating and enhancing the campaign in some way?
Doing this also lets your campaign gather a sufficient amount of data for predictive models to learn and iterate from over time. On Day 1, the model has no historical data to work with. It hasn’t observed conversions or learned the nuances of who clicks and who converts.
Models update every day, incorporating new data to refine predictions. This iterative process improves targeting and bidding strategies over time. As the campaign progresses, the model gathers valuable data on post-click and post-view conversions, enabling more accurate predictions and optimizations.
3. Recognize a trend versus a fluke
Collecting data over time allows you to observe patterns that may not be apparent in a single day’s results. A true trend should reveal how your campaign evolves and performs under various conditions.
Giving yourself more time also helps to distinguish the flukes – random or coincidental results that might initially look worthy of taking immediate action. If your campaign starts off amid the holiday season, for example, performance could look a lot different than at other times of the year.
A longer window of time lets anomalies average out, leading to more reliable and representative results. It gives you visibility into what drives consistent performance, making your findings more trustworthy to inform strategic decisions.
Conclusion: How to strike the right balance between data and decisions
None of this is to suggest you should run campaigns indefinitely without scrutiny. The key is to find the right balance:
- Consult with your Customer Success Manager to understand the optimal data collection period for your specific campaign goals.
- Set checkpoints to review performance trends rather than making decisions based on single-day snapshots.
- Be prepared to give your campaign enough time to gather statistically significant data before making major adjustments.
It’s hard to be patient sometimes, but in performance marketing it comes with a significant payoff – developing an informed strategy to achieve ROAS goals. Connect with us to get started on optimizing your campaigns!