Last updated 30 October, 2021.
Whenever you talk about growth marketing, the discussions about experimentation are waiting just behind the next corner. If you don't feel like reading, watch our webinar on the same topic 👇.
But what is this experimentation? It’s a crucial part of growth marketing, that’s for sure, and a critical puzzle piece in your process for growth.
One of the hallmarks of experimentation (for me personally) is that while it's always quite goal-oriented, your most important objective is to learn something.
Here’s an example from real life: One of my customers was excited to hear some ad experiments weren’t working. Why? Because this way, we learnt that the channel we tested wasn’t working for our specific target audience.
And in this lies the spirit of experimentation.
The quintessence of experimentation is to test different things and learn what works best. Or what doesn’t work at all. Or what works somewhat. You get the idea!
This is also one of the secret superpowers of a growth marketer — or what distinguishes a growth marketer from something else: it’s the growth mindset.
It’s not to say that those who aren’t a growth marketer can’t have it. But while you can perform marketing tasks without a growth mindset, you cannot do growth marketing without it.
A growth mindset is all about learning. While it might sound like something closely related to “Good vibes only,” it simply means seeing things as learning opportunities rather than failures.
And this is the mindset we want to implement into all marketing strategies. ❤️
(Btw, if you're interested in this topic, stay tuned for more and subscribe to our Advance Insider newsletter to make sure you won't miss any future posts! 👇)
Where did the shift towards experimentation come from?
Well, I’ll be the first to admit that I am not 100% sure where this comes from.
The shift is simple evolution: A combination of many things that has been boosted especially by things turning digital, everything becoming more measurable, and organizational silos coming down.
However, it’s something we've seen with the advent of growth marketing (and growth hacking, for that matter.) A few years back, marketing would plan the next year all at once, then maybe plan different campaigns, then run these rather straightforwardly.
At that point, inbound marketing became popular (thanks, HubSpot — we love you,) and marketing was highly centered around content.
I’ve heard stories where it took several months to plan one marketing campaign from an e-book to ads to email promotion — and infographics.
I believe three things happened that shifted this approach towards experimentation for business to business growth companies:
🤔 More data became available. Marketing is a function that has a lot of data at its disposal. Having this data invites us to analyze, optimize and adjust it accordingly.
🤔 Marketing and sales were drifting further apart than ever. The e-books marketing teams were producing were just not bringing in high-quality leads for sales.
🤔 Marketing budgets were downgraded across organizations. Let’s face it, it’s typically the first budget that gets cut — and when these things happen, marketing teams need to do more with less.
Simultaneously, you had tech and software companies from Silicon Valley taking the world by storm, quite literally hacking growth, with smart marketing activities.
We all know the examples: Hotmail’s “invite a friend”, Dropbox’s “share with a friend for more space”, and the likes. These kinds of activities set up new standards for how marketing had to operate to meet these new kinds of challenges.
Typical learning objectives
Marketing teams from small and large organizations transformed themselves into growth marketing teams and set out to experiment.
To recap: in experimentation, it’s more like a discovery approach. It’s about answering questions such as:
⁉️ Does this channel work better compared to others?
⁉️ What kind of messaging will be interesting to these specific customers?
⁉️ What kind of other content will they be interested in, or more importantly: What kind of content will address their specific needs?
Growth marketing experiments are designed to find out what works best. When you learn what works even a little bit, you can narrow that down to what has worked the best and do more of that. Fun, right?
Lead generation is just one simple example.
You can (and should) approach experimentation from the whole pirate growth funnel. But bear (🐻) in mind that you need to make sure to define your goals and metrics along each stage of the funnel, as they are likely to differ quite a bit.
Advertising channels are an easy way to experiment — they are one tactic if you will. Often advertising is used because it’s relatively easy to get to the volumes you need.
Let’s look at an example. You are working on renewing your brand messaging, and your homepage has 1,000 views per month.
Because that’s a relatively low number, testing two different copies for your home page might not be the most conclusive experiment on its own.
However, creating ads with new messages and running them for a month could be a better option. This way, you probably get well over 1,000 impressions, and those might offer you more insight into what works.
Four things to keep in mind when moving towards experimentation in your B2B marketing
Let’s face it; it’s not rocket science.
There are, however, some mistakes I see really frequently with experiments. I want to leave you with these four tips so you can start experimenting right away:
Stick to one KPI or goal to test. Define one thing that you want (e.g. leads or lead to customer conversion). Define one metric to focus on.
Focus on specific things. For example, if you have tons of segments, focus on one segment. Define a small enough box to throw ideas into.
Don’t get distracted. And start small.
2. Top of the funnel
Start from TOFU experiments where the sample size is big enough. Think acquisition channels and your top landing pages.
In the B2B space, there’s often not enough data down the funnel, and after becoming a customer. (With your existing customers you can make bigger bets at once, and rely more on qualitative data.)
Advertising is an easy way to start, as you can get a lot of impressions and quickly see who clicks most on ads. Once again, not rocket science but something that you can learn from.
This one’s important. Make sure you follow a framework for experimentation.
Otherwise, you are not experimenting but just aimlessly doing different things. Start by defining your goal, focus, and hypothesis. Add why you think it would work, and refer to what you know from your customers.
Define metrics that will suggest for your hypothesis to pan out, test, and record your findings. If you have many ideas, I’d recommend using something like the ICE framework to prioritize your experiments.
Start with something small and simple. And on that note, if anyone ever tells you that you need to run a lot of experiments, ignore them. It doesn’t matter how many experiments you run; it matters how well you run them.
A lot of experiments can correlate with more learning, but first you need to nail down the process.
In B2B and small markets, there is only so much audience and website traffic you can play with. This means that focusing on quality makes sense.
Especially in the B2B SaaS industry, there is a limit of what is sensical to test.
There are a million things to test with consumers, and the volumes at larger B2C companies are likely high, but if your SaaS product caters to some niche at small B2B companies, you probably won’t get up to volumes that warrant such huge experiment quantities.
And of course, if you have the necessary resources is a whole different question.
So, to recap:
✔️ Define your goal and focus area (eg. new trial users from a specific market)
✔️ Do your customer research and data analysis (what has worked, what could work)
✔️ Proceed with continuous experimentation:
- Hypothesis: Refer to research, analysis, previous experiments
- Define experiments and test
- Prioritize experiments that are fast to run, have a strong likelihood to succeed, deliver enough data to learn from, and are repeatable and scalable
- Track, learn, run another test
4. Don’t give up.
The biggest risk is the wrong mindset. Don’t set out to test something and then abandon the whole thing the instant you see it doesn’t bring results.
Always, always look at this as a chance to find out more about your potential customers and why something may not have worked.
Here are some more ideas for you to borrow to start experimenting:
KPI examples: Brand awareness, branded searches, engagement
📈 Test different pieces of content in advertising to see what resonates with the target audience. Which blogs are being read? What videos are being watched? If people consume content, they’re more likely to remember you — thus growing your brand awareness.
📈 Take your messaging and content and run continuous brand awareness campaigns. See if it results in more searches for your brand name. A good indicator of brand awareness, and a SEO juice booster!
KPI examples: Trials, qualified leads, cost per conversion
📈 Test different content types such as e-books, video content, and infographics to see which is the most interesting one for your audience
📈 Test new messages with ads and see which ones resonate with your audience
KPI examples: % of users that experience the core value of your product
📈 Carry out customer research, and check your onboarding from the customer perspective. And then fix it. This is important, and it can be a larger-scale experiment.
📈 Use a product analytics software and segment your users to different groups, see if they have different Jobs to be Done, and make personalized onboarding flows. Like Intercom did.
KPIs: Customer Acquisition Cost, Customer Lifetime Value
📈 Reverse engineer the path from a stranger to a customer: what content did they view? Which channels did they use? Draw new hypotheses and experiments out of these to see if you can boost your customer acquisition efforts.
KPIs: # of new customers coming from referrals, NPS
📈 If your product enables it, add different “refer a friend” options inside the product and see which ones generate referrals.
KPIs: Churn (# of customers, or monetary value)
📈 Help users get the most out of your SaaS product. Analyze product usage with product analytics. Make hypotheses about product changes, test and measure them. For example: “adding a product tour will increase product usage”.
And there you have it. Better start experimenting!
If this resonates with you, check out our growth marketing services. We’d love to hear from you! 😍
And in case you're looking for something to read next, you can check out our short guide to building a B2B SaaS go-to-market strategy 👇