Subscribe to OUR newsletter
Sign up and do more with our non-profit/fundraising tips and helpful resources.
Every organization, be it an e-commerce store selling novelty t-shirts or a nonprofit helping to solve climate change, has a set of desired actions they’d like to have occur on their website.
In some cases, these desired actions are simple to explain – for instance, it’s fairly clear that on an e-commerce website, the “conversion” event occurs when someone buys a product.
However, in some cases, it’s not so clear. What if you run a non-profit site? Certainly, there are some goals you’ll want to achieve with the website, including informing donors, education the population on your cause, and updating those who have a stake in your cause with recent news.
Despite these goals being somewhat less “obvious” than someone adding a t-shirt to their shopping cart, conversion rate optimization can still help you.
Conversion rate optimization (CRO) is the systematic process of increasing the percentage of website visitors who take a desired action. This action could be one of many things, from the common example of filling out an email sign up form to something more obscure, like reducing the amount of unsubscribes (un-conversion rate optimization, that one is coined).
No matter what action you’re trying to make happen more often, the point is that CRO is a process and it leans on data to drive decisions, rather than top-down opinionated decision making.
Normally, in CRO, we use controlled experiments, or A/B tests, to make these decisions. We also rely on other tools to inform our decisions, such as digital analytics, qualitative surveys, and user tests.
CRO, it’s important to note, is not a tactic or a set of tactics. It’s a process and a different way of looking at designing websites and marketing materials. It brings the decision outside of the conference room and into the hands of actual website visitors. In my mind, it’s the most customer-centric way of doing business.
So if CRO is all about increasing the rate at which visitors take a desired action, it follows that we first have to define what the desired action is on our website.
That’s where things get a bit complicated. In most cases – not all, but in many – there are many desired actions on a website. In e-commerce, it may be as simple as buying items. In nonprofits, you have many goals and stakeholders to address.
There are usually two large problems with organizations set up like this:
In the first case, you simply need to take some time and have a tough but valuable conversation: why are we online? What do we hope to accomplish?
I like a thought experiment that the folks from the Digital Analytics Power Hour podcast put forth. As they put it:
“Imagine you wake up in a world where you don’t have a website. In this world, what would be your argument for creating one? What would you tell your boss to get them to invest in a digital presence?”
This type of exercise is fruitful for any type of organization, by the way.
For example, a SaaS website exists to sell and deliver software. A government agency might exist to provide accurate information to individuals trying to find their local DMV.
A nonprofit might have a website to promote their cause to the world that doesn’t already know about them.
They may want to drive net new donations. They may want to provide news for current donors and stakeholders. It might be a combination of all of those things, and maybe there is a hierarchy of goals, where one takes priority over the others.
These are things to discuss and define, as there isn’t a one-size fits all answer to every organization.
When you do define a goal or a set of goals (though hopefully one in particular that you hope to move the needle on), you can move forward on the conversion optimization process.
Almost every conversion rate optimization process follows, at least at the very highest level, some combination of the following steps:
First, you do research to find what problems exist on your site. This assumes, of course, you’ve set the goals you want to hit and what metrics you care about. Then, you can work backwards, and ask, “what parts of my website are currently not contributing to our goals or are underperforming?”
Some of this, you’ll have quick ideas about. You may immediately be struck with the realization that, “oh, our lead generation forms are like 14 fields long. That can’t be good.”
Some of this, however, you’ll have to search for and find by pleasant surprise via conversion research. Some great tools to dig up conversion insights include:
You’ll find which methods you like best at your organization. Personally, I love digital analytics to find what the problems are and where they exist, and then running user tests and customer surveys to find some potential hypotheses as to why they’re occurring and how we may fix them.
Also, there’s a lightweight version of each of these tools and then there’s a whole rabbit hole you can dive into. With Google Analytics, you may simply know how to view landing page conversion rates by traffic source, or you may be able to run advanced path analysis. Of course it’s better to know more, but you can still learn a lot with limited skills in some of these areas.
Some of my top research tools include, outside of Google Analytics:
All of this research then transfers into the hypothesis and ideation process. Essentially, you have a pile of research and insights; how do you change those into action steps that you can actually test out to see if they improve things?
This is where you start to hear all the talk about forming an “experiment hypothesis.”
What’s a hypothesis? It’s a proposed and falsifiable statement about the way you believe the world to work. In our case of online marketing and optimization, it’s usually a belief about our users and their interaction with our website experience, such as:
We believe people want to donate to climate action but simply don’t understand the messaging on our landing page, because we’ve consistently seen a high volume of questions regarding our mission on our donations landing page.
From there, you can come up with different ways to test that hypothesis (from this example, I can easily see some copy tests, changing the design and layout, etc.)
Next, you’ll want to prioritize your ideas. Usually, you’ll do this through some sort of scorecard made up of important variables like impact (how much upside you can expect if the idea works) and ease (how much it will cost or how long/complicated it will be to implement).
Finally, we run our test! While the ideal implementation would be to run an A/B test, a controlled experiment that pits the current version against your new idea(s), sometimes that’s not possible. You could be limited by many reasons:
The most understandable of these reasons is the lack of traffic. Math is math, and you need a certain sample size in order to actually run valid experiments. If you don’t have the traffic, you can’t run valid tests (though you can still optimize your site! User testing and qualitative research are your friends).
If the case is a lack of time, you can usually do one of two things:
Bandits are a bit complex, so I’d stick to simple A/B tests in most cases, and if you really need to rush something to production, make sure to at least run some user tests and usability tests to see that it works functionally.
As for lack of resources, if that’s the case, I totally understand. We’ve all got limited resources. In that case, it may be best to over index on “ease” when you’re prioritizing tests. No sense in trying to launch a super complex, high cost test if you don’t have the resources to see it through.
Running an A/B test is easy – we have the software solutions today. But running a proper A/B test? That’s difficult. After all, we’re dealing with statistics and a lot of interesting online variables (bot traffic, flicker effect, external validity factors, just to name a few). The best answer I can give is find someone who knows what they’re doing to at least walk you through your first tests. Running bad tests is probably worse than not running any at all (at least you have the epistemic humility to say “I don’t know” if you don’t have the data of an A/B test to back you up).
This could be a consultant or even an employee who is stats savvy at your company, but it really does help to have an experienced eye to QA your test. You can (and should) also educate yourself if you plan on running many tests. Luckily, there are many resources available, including:
Another thing with all of this: never stop learning and improving. Your site isn’t static, and neither are your visitors. In addition, there are always more things you can learn about testing, CRO, human behavior itself. It’s a deep rabbit hole, but one well worth diving into.
Conversion rate optimization and A/B testing aren’t only for Amazon, Google, and e-commerce sites. You can use them to increase donations on your website, improve the user experience of any given page, or even provide clarity and resonance to your value proposition and mission statement.
Unfortunately, it’s no quick fix; it’s really a mindset, one where you approach decisions empirically, encourage innovation and reduce risk via experimentation, and realize that the road to success is paved with small and incremental wins over time.
To learn more about how Paperless can help your organization, request a demo today!