How to Use a Test-and-Learn Approach to Come Out Ahead of a Crisis
Under normal circumstances, most marketers would agree that taking a test-and-learn approach to build actionable insights for improving advertising effectiveness is a smart move. But now, testing and learning through marketing experiments might initially seem like a luxury.
How can they justify diverting time away from standard campaign management tasks to build insight? It might seem wise to just buckle down and keep things simple, but to think about it another way: If you're not testing, your competitors most likely are.
Who Has Any Historical Data to Guide Them Right Now?
During a non-global crisis period, most digital marketers can take a quick glance at their weekly or monthly KPIs and gain an immediate understanding of how things are working, asking themselves the following questions:
- Are CPCs high or low?
- Is impression volume rising or falling?
- Do CTRs look normal?
An experienced omnichannel marketer working for years with an account may not need a marketing experiment to determine these basic trends.
However, right now, historical data won’t really be worth much to anyone. There’s nothing comparable in our recent past to today’s environment. Directionally? Maybe. But true navigation into what is and is not working requires real insight.
While CPMs might be really low compared to a few months ago or last year at this time, do they correlate to more ROI if consumers are holding on to their cash before they spend on anything non-essential? What about channel allocation? Should marketers cut budgets or push forward? Should they spend more in search or put that incremental spend in social or elsewhere?
Certainly, messaging is a key area to be revisited during a time of significant change, but what messages? Which ones are working? Which ones need to be pulled immediately because they are unintentionally causing negative sentiment?
These are the kinds of questions marketers need answers to, and unfortunately, data from pre-COVID campaigns likely will not help right now.
Data-Less Decisions Are a Guess at Best
Hopefully this pandemic will be over soon, and we’ll all be able to return to our normal routines. Yet, the truth is that there are always market fluctuations. While they won’t be as big as those during the pandemic, there are always forces at work impacting marketers’ campaign performance. It could be competitors ramping up their marketing activity, macro-trends like politics, supply chain friction, or a million other invisible variables that make advertising metrics go up or down.
A test-and-learn approach offers a reliable method for truly understanding the way various marketing elements — channels, campaigns, ads, bids, etc. — move the bottom line.
Many organizations might be tempted to postpone testing until business gets back to normal – and it’s true that current results will likely require retesting once the pandemic is over. However, for most marketers under pressure to make the right decisions, testing can serve as the best solution right now to take data-driven decision making to the next level.
After all, according to Gartner, organizations that significantly outperform their competitors are almost twice as likely to make testing and experimentation a marketing priority. That’s why it’s imperative to execute testing continuously, as situations change over time — what was true yesterday might not be true today.
The present situation provides marketers an opportunity to establish a test-and-learn approach with their teams, learn the tools and testing methodologies, and build internal processes to be ready to hit the ground running once things open up again and be better prepared to respond to future disruptions and changes in consumer behavior.
Leveraging Marketing Experiments for Data-Driven Decisions
The surest way for marketers to test, validate, and calibrate their data-driven decision process is through marketing experiments. Ideally, an ongoing test-and-learn culture is the best approach, and marketing experiments could eventually become part of every marketing team’s job.
Understanding whether a decision was good or bad is just one benefit that marketing experiments can help accomplish, but whatever decisions they seek to make, how do marketers know which data to use to base these decisions on?
For example, for a digital marketer analyzing an ad to decide whether or not to pause it or let it keep running, a few of the available metrics might be: costs, CTR, and conversions/conversion rate.
One marketer may compare CTRs of other ads and decide it’s not working well and pause it. Another marketer might look at the conversion rate and keep the ad running. Still another may look at both the CTR and the conversion rate and come up with another course of action.
What about time range? Should a marketer look at just the previous 30 days? Year-over-year comparisons? Each metric changes based on the time range chosen and could absolutely push a decision in different directions.
Global freelancing platform Upwork sought to improve business planning by understanding the effectiveness of marketing campaigns for new audience growth programs. Upwork performed incrementality testing to discover the impact of SEM brand keyword campaigns, a first step in tightening budget projections that would determine the incremental value of an entire keyword category and provide greater confidence in evaluating campaigns across other keyword types.
As a result, brand SEM campaigns proved to provide incremental lift/value. Using precise metrics surfaced by testing, Upwork effectively realigned channel bid targets to effectively harness brand keywords and work better within customer lifetime value models.
A continual test-and-learn approach with marketing experiments at the foundation can also substantiate which datasets drive the best and most accurate decision-making.
A Test-and-Learn Approach Requires the Right Technology
With new innovations in big data and cloud hosting platforms, modern tools can make testing that was once very complex, expensive, and slow an easy process that’s highly affordable, offering fast results that any marketer can run.
Embracing a test-and-learn mindset creates an advantage for organizations, with the right technology providing the ability to quickly test assumptions they might not have attempted when it meant bringing in a process-heavy analytics team to accomplish. Standardizing marketing experiments so everyone on the team can run tests opens up more analytics-focused decision-making, allowing teams to find answers to inspirational hunches they may not have normally felt were within their realm to figure out.
Equipped with the right solutions, marketers can unlock smarter planning, better budgeting and stronger alignment to business objectives by measuring and monitoring the true business impact of their marketing programs.
Marketers must be able to count on their decision-making to be nearly flawless to compete in today’s global marketplace, especially during the COVID-19 crisis. With brands increasing their skill at executing campaigns across all touchpoints, simply being available to customers isn’t enough to differentiate their business. Opportunities come to those with more insights to test, with more discoveries to differentiate themselves in a down economy.
Moti Radomski is VP of Product for Kenshoo, a global leader in marketing technology. He has over 20 years of experience mixing business, product and technology across business intelligence, analytics, SaaS software, ecommerce, and online advertising. Moti can be reached at firstname.lastname@example.org.