Tuesday 7 February 2017

How to Build a Process for Growth Experiments

Growth Experiments Header STock Photo.pngWhen you work inside a business, it can feel like all anyone talks about is growth. Revenue growth! Customer acquisition! Leads goals! Increase our traffic! The list goes on.

While the idea of “growth” -- and all of the hacks, tips, and tricks that go along with the term -- can often feel a bit buzzy and overwhelming, real growth actually is actually the result of well-established processes and an internal emphasis on experimentation and testing.

Brian Balfour, former VP of Product at HubSpot and a thought leader in the world of growth marketing, is an expert when it comes to building a process for experiments that enable growth overtime. While he emphasizes that there isn't a one-size-fits-all solution for growing your business, he does have some pro tips when it comes to actually building a process for growth experiments. And that's what will ultimately move the needle for your company.

We'll take you through the steps of a building a sustainable process for growth experiments based on Balfour's tips. Once you establish an experimentation process that works for your team, that buzzy growth term might seem attainable after all.

And once you get there, our new 2017 Marketing Experiments Calendar can help you build a backlog of ideas, design and prioritize your experiments, and record your results all in one place.

How to Build a Process for Growth Experiments

To grow, you have find ways to experiment. The problem with that is finding time to do that in the right way, especially when you also have multiple other tasks and ideas. That’s why building a sustainable process for recording your ideas and implementing them as experiments is essential. By figuring out these steps, you’ll enable your team to effectively write down ideas, prioritize experiments, and put them into practice.

But where do you begin?

1) Start with the scientific method.

Ah, yes. We’re back to Week 1 of any science class you've ever taken. As it turns out, you really do use some of that elementary school information in real life.

 

Scientific Method Source: Science Buddies

When it comes to building a process for experimentation, it comes down to following the scientific method. Create a way to brainstorm ideas, construct hypotheses from them, implement them, and prioritize your ending analysis and conclusions.

When experimenting for the sake of growth, building this process for your team isn’t a one-and-done deal. Instead, according to Balfour, a good process has to be:

  1. "Scalable
  2. Predictable
  3. Repeatable"

In other words, it’s not enough to just say that the scientific method is your process. Instead, use the scientific method as template to structure your internal process. 

2) Always have a goal in mind.

Once you have your process down, the next step is to set goals, which might sound simple in theory. But the goals of any given experiment can't just be "growth" -- because growth along isn't measurable. Instead, every experiment you run should have an actionable and testable outcome, which means you have to start small.

Let’s break it down with an example:

  • What's the overall goal? Revenue growth.
  • How are you going to drive more revenue? By acquiring more leads.
  • How will you do that? Increase the conversion rate of my lowest-converting lead generation campaign.
  • What is your campaign-specific goal? Increase conversions on one landing page by >5% and find statistically significant results.
  • How will you implement an experiment to meet that goal? A/B test the title of the lead generation offer.

You can see from th example that while increasing your revenue is an admirable high-level goal, it’s not actually actionable unless you break it down into actionable experiments. And tying a metric to your experiment -- in this case, increasing conversions by greater than 5% on one landing page -- is what makes it actionable.

If you have trouble setting actionable goals for your experiments, start with that high-level goal and work your way down an incremental path, until you have something that you can tie to a metric.

3) Encourage and enable brainstorming and hypothesis.

Building a culture of experimentation within your team requires the encouragement and motivation to always be brainstorming. That's why building a mechanism for tracking and recording those brainstorms is essential for any good growth experiment process.

Whether it’s an Excel sheet, Google Doc, Trello board, internal wiki page, or any other tool your team typically uses for internal documentation, make sure there’s room for any and all brainstorming.

Trello Brainstorm Doc

Still, it’s not enough to just throw raw ideas into a document somewhere. Instead, you have to require your team to take those brainstorm ideas and turn them into actual hypotheses based on research.

Here’s an example of what this looks like in practice:

  • Idea from brainstorm doc: Use a pop-up modal instead of a in-line CTA to increase conversions on our blog posts.
  • Research: Our in-line CTAs are only converting at a rate below 1%. If we use a slide-in or pop-up CTA, we could make the CTA more visible to boost conversions.
  • Hypothesis: By using pop-up modal CTAs, we will increase our conversions from a blog post to a landing page by >2%, with statistically significant results.

4) Figure out a way to prioritize what comes first.

Once you’ve enabled brainstorming and hypothesis-writing on your team, you’ll might find that you have more ideas than you know what to do with. That's not a bad thing -- remember that growth occurs because of many moving pieces. The more things you have to test, the more you’ll be able to grow over time.

That said, with hundreds of ideas being thrown around, you'll have to figure out a way to actually prioritize which experiments you put resources into first, and why. While there are many ways to do that, at HubSpot, we often use the PIE method - “Priority, Importance, Ease.”

We execute this method by compiling all of our ideas into one document and then rate, on a scale of 1-10, each project in the categories of Priority, Importance, and Ease. Then, we take the totals for each category, and divide those numbers by three. That gives us a PIE score to rank each individual project, and we can better determine which experiments or projects to work on first.

PIE Score

No matter what process you use to prioritize experiments, always remember to focus on the ideas that you think will have the biggest impact. That can help you prove the value of your experiments to your team and company, and drive that culture of experimentation forward.

5) Take time to design and implement effective experiments.

Now comes the fun, but most challenging part -- implementing your experiments.

Whether you’re performing an A/B test on a landing page, analyzing creative content on your social media ads, or looking at the full conversion path of your lead generation offers, take the time to fully think through the details of your experiment. Here are a few tips to keep in mind:

  • Test one thing at a time. The more variables you have in one experiment, the less meaningful your results will be. If you test one landing page design against a completely different landing page without any experiments in between, your results won’t tell you much about what actually worked in that experiment. Instead, if you test one variable at a time -- like the title, imagery, colors, fonts, and module placement -- you’ll be able to learn from each experiment which variable is actually causing a difference in results.

ab testing_growth.jpg

Source: HubSpot Academy

  • Set up a way to measure your results before the experiment begins. Many of your experiments will require a tool to help you actually perform it. Whether you’re using a content management system or designing an experiment by hand, understand how you’re going to measure and assign meaning to your results before you get started.
  • Set a time frame. How long will you let your experiment run before you call the results -- one day? Two weeks? Until the results become statistically significant? Setting a time frame will not only help make your experiments more accurate and consistent, but also, will help you track results without letting the experiment fall to the wayside. Once you’ve set a time frame, set a reminder for yourself to analyze the results on the day the experiment is complete.
  • Learn from experience. Use the findings of your experiments to come up with new ideas for future ones. Your experiments don’t have to be a one-time engagement. Good growth is built on a culture of constant experimentation.

6) Always analyze your experiments.

When the timer goes off and your experiment is complete, it's critical that you actually analyze and record the results and use them in practice. But it's not as simple as learning if your hypothesis was correct or not. Instead, the analysis needs to answer the following questions:

  1. Was my hypothesis correct or incorrect?
  2. Are the results of my test statistically significant?
  3. What do my results mean?
  4. What could/should I have done differently in this experiment?
  5. How can I use these results going forward?
  6. What should my next experiment be to build on this finding?

Building a culture of experimentation isn't just about getting a “pass” or “fail” answer on an experiment. Instead, true analysis happens when you think about what the results say about your efforts, and how you can use what you learned to move forward with new experiments.

7) Share and systematize your learnings.

Before you repeat this process in your next experiment, it's critical to share the results of your experiment with your colleagues. Sharing information internally is a great way to both encourage transparency and generate feedback -- which can help you build upon a history of experimentation.

At HubSpot, we use an internal wiki where we post experiment results, in addition to all kinds of internal information for employees. It’s a great way to keep everyone in the loop on what’s going on, as well as get feedback from colleagues who might be interested in your findings.

You can also keep a running document of your experiments, so that the findings are easily accessible in one place. For example, you could create a document where you track all of your A/B tests from your landing pages or all of your tests on CTAs.

Process for Systemizing Learnings

Build on What Others Have Done Before You

Now you know how to build a growth process that will be scalable, predictable, and repeatable. But if you’re like me, you probably want a document to use as a starting point,  and build from there.

Luckily, we've created a Growth Marketing Experiment Template based on this process that you can use and customize for your own internal process. Want your own free copy? Download these free Growth Marketing Experiments Templates now.

What does your experimentation process look like? Let us know in the comments.

Marketing Experiment Templates



from HubSpot Marketing Blog https://blog.hubspot.com/marketing/growth-hacking-experiments-processes

No comments:

Post a Comment