Introduction
Two years ago, before starting our first tech startup, Viral Loops, we built one of the first growth marketing agencies, GrowthRocks.
I’m starting to lose count of the companies we’ve worked with, but almost all of them wanted to learn the process we followed for our internal needs.
So, what we do, is simply a sum of Sean Ellis’ high tempo testing method, Brian Balfour’s growth machine and our amazing growth team testing like lunatics.
High tempo testing method
The community of GrowthHackers.com is an excellent example of how high tempo testing can skyrocket your results.
As it’s a community, GrowthHacker.com main key metric in the number of monthly active users.
They managed to nearly double their monthly active users in 11 weeks from the moment they started testing in high tempo.
As Sean Ellis points, it’s important to run as many experiments as possible in a given timeframe, as you will get more insights, in a quicker manner, about what drives growth for your product.
In order to discover the right experiment you first have to explore, which translates to running more and more experiments.
As you can see, we start with brainstorming and documenting new experiments, which then we prioritise and we decide which of them we are going to implement.
The next steps are to implement, run and then analyse. documenting everything is important, as at the end of the day you can systemize easily successful experiments, making them that way a part of your marketing approach.
Note:
When you’re in the ideation step of your experiments you should keep your focus on inputs rather than outputs.
For example, if we’re trying to increase your Monthly Recurring Revenue, you should build experiments that will influence the number of free trials and your conversion rate.
So, increasing Conversion Rate is one of our OKR’s, for which we have a specific document.
The OBJECTIVE KEY RESULT document is the place to store the key results you want to hit in the given timeframe of each experiment.
Instead of setting just one target result, we like to set 3, scaled based on how hard is to reach the goal.
In the particular example, you see that the experiment is to increase our monthly conversion rate.
As a more long shot result, we set a 15% monthly conversion rate in a 60 day period.
We’ve also set 2 more tiers, which are more attainable.
Make sure to keep in mind that each experiment should run from 1 to 3 months. That gives you enough time to acquire a reliable statistical depth unless you have tons of traffic.
In that case, you’ll be able to test in higher tempo, which will allow you to bring results faster.
Generally, the higher your testing mode, the more insight you’ll be able to get. You have to build your experiment’s hypothesis, using a logical expression IF THIS, THEN THAT, BECAUSE THIS.
For example,
“If we increase the traffic by 100%, then the conversion rate will double because the checkout process is already performing in the desired way.”
Also, you can note down the exact steps of the experiment’s design.
For example,
Step1. Build a landing page
Step2. Send a dedicated email newsletter
Step3. Run a FB ad etc.
Adding an estimation of the man-days that each aspect of the experiment needs, will help you prioritize better the experiments you should run. Running the wrong experiment can really spoil the momentum you’re trying to build.
E.g. 0.5 man-day in the landing page development.
Also, it’s wise to keep track of your key metrics along with some action items for the next sprint.
Instead of dreaming big, you have to be pretty down to earth about your experiment’s hypothesis.
Aim for drastic things that will bring meaningful results.
Adding the ability to checkout without subscribing is by far a more meaningful test than changing the checkout button.
Last, there should be WEEKLY LEARNINGS document for keeping the things you learn every week as far as it concerns each experiment.
This is where, you will store your learning and the data that supports it, plus your actions for the upcoming week.
Keeping a consistent documentation will help you establish a knowledge base for your team in order to avoid mistakes in similar experiments.
After every experiment ends, we store our learning using a Wiki software. We currently use TikiWiki, but you should find the one that will work best for you.
Our typical week
The week starts by summing up what we’ve learned during the previous week and how to leverage those learning to our own sake.
Prioritising new experiments also belong to Mondays, and since everything’s set and clear, the team is entering in execution mode which lasts to the very end of the week.
Tuesdays are for discussing technical stuff and doubling down on tasks.
Wednesdays are quite office days, as everyone is trying to nail it like a boss.
Then on Thursday, we take a break from breaking stuff and we put together all the ideas that we have for the next week.
Now, we all know that Fridays are not exactly what one would call a productive day. that’s the reason we try to keep executing at a low level and instead focus of recapping the things we’ve learned so far.
I must make clear that we are not robots. We keep the cyborg side of ourselves only for executing. We also have food parties, we talk nonsense and harsh jokes to each other and play with our mascot puppy.
The living is easy!
You probably think that what I told you so far, are actually harder than I make them sound.
You are damn right. But you can use our stack to make it easier for you.
Tools we use.
We use GrowthHackers projects for keeping an eye on our experiments–both running, backlog and pipeline.
Then we have Paper by Dropbox (which rocked my socks off) for any kind of documentation, such as learnings or metrics.
Trello is our to-go for task allocation and management, and finally, slack (Geekbot) for internal team communication and sending memes.
Real life example
To better understand the process, let’s see how this process can apply to real life.
In Viral Loops we are continuously looking for new ways to increase our e-commerce customers, organically and without ads.
So let’s dig deeper.
Step 1
Firstly, we came up with a few ideas and we added them to the experiment backlog. Such ideas were:
1. Write guest posts in big e-commerce blogs.
2. Put together a super actionable e-Book about e-commerce growth
3. Create an e-Book in collaboration with a company that provides necessary an e-commerce tool and can share it with its customer base.
Step 2
After prioritising these ideas, we decided to focus & execute #3 because it seemed more promising to get some new e-commerce customers.
That’s why we picked Contact Pigeon, an email marketing automation platform for e-commerce stores.
The e-Book is titled Let the Ca$h Flow: A Guide Full of eCommerce Growth Tactics.
Check it out ???? here.
Step 3
This is where you have to understand if the experiment worked or not. In order to do that you have to measure it first. These are the metrics we decided to track:
– #number of e-Book Downloads
– New sign-ups that came from downloading the e-Book
Step 4
We documented exactly what we did. From how we wrote the e-Book, what format we used along with the marketing channels we used to promote it.
Conclusion
Every company is different. All of us are facing marketing issues and we have to take advantage of chances that come up along the way.
What is going to help you survive is your team and your process.
These two will help you surpass every obstacle on the way to success and growth.
What process do you follow? What would you do to grow your numbers?