How to Start an Optimisation Program

So you’ve decided you want to start a web optimisation program. Well done! If you’ve got a website, this is definitely the right direction for you to move in. Being more data-driven with your decision-making can only be of benefit to your business. While plenty of people made it through 2020 and now 2021 by “ignoring the numbers,” this is never an actually beneficial technique. And if you’re looking for inspiration around experimentation, you can see just a small sample of the results we’ve achieved with our clients at LeanConvert through A/B and multivariate testing here.

As with any new business practice, however, getting set up with experimentation is easier said than done. Below, we’re outlining some of the crucial considerations for you to decide on before diving into experimentation.

Is the Timing Right?

The simple answer is (usually!): yes. It’s almost never the wrong time to start experimenting. If you as a business aren’t A/B testing, you’re either going with your gut (which means you could consistently be making the wrong choices for your customers), or you’re experimenting by looking at numbers after the fact (without a control group). No matter how fine-tuned you feel your instincts may be, they’ll never be as accurate as empirical data.

And if you are just guessing, you’ve already invested time and money into hard coding changes. This after-the-fact approach doesn’t account for seasonality—what if the market happens to pick up just as you launch a feature? The numbers will look positive, but without a proper control group, it’s impossible to draw real conclusions. Effectively, you have no idea what’s really going on with your CX.

The sooner you start, the better. Meaningful insights take time to gather, and experimenting allows you to try riskier things that you wouldn’t just implement otherwise. There is an opportunity cost to not having an optimisation program. The longer you take to get started, the longer it takes to start realising value from riskier but well-performing features. Why continue to leave so much unclaimed value on the table?

Any time spent not testing means risking heading further and further in a direction that’s not the right path for your customers.

What Are Your Goals?

It’s important to be clear on the strategic objectives behind the testing program. Some companies will start off with objectives such as “run X tests per month” or even something as vague as “do more personalisation.”

Thinking in this way is putting the cart before the horse. Optimisation and personalisation are means for achieving your objectives, not objectives within themselves. This misunderstanding is one of the most common (and dangerous) faults we encounter.

Strategic business goals must come first in order to ensure that you are only doing what is needed and valuable. If a test is easy to build but it doesn’t help with your primary business goals, it shouldn’t be run.

Putting optimisation, personalisation, or tool usage first and your business objectives second will naturally mean doing things that are needlessly complicated for the sake of using a complex tool, or to prove that you’re doing personalisation. It’s easy to build a bunch of simple and poorly thought-through tests to reach a monthly quota, but are they actually worth doing?

Having a focused approach here is better than a scattergun approach. It’s important to create finite and clear targets. For example, perhaps you’re aiming for a 3% increase in the new customer sign-up rate over the next 3 months.

These optimisation goals should align with your larger business objectives (such as focusing on customer acquisition rather than increasing the lifetime value of returning customers). Setting goals once per quarter is a good practice to make sure you are not changing directions too often (every month, for example) while still remaining flexible. It’s easy to get overly excited or curious about experiments that don’t actually align with your overarching objectives.

What Resources Do You Have?

So you’ve figured out your goals for the first quarter of experimentation. Now the question is, how do you get there? This depends largely on the resources you’re working with.

Firstly, one of the obvious cores of a successful optimisation program is the testing tools themselves. Needless to say, there is no “best,” one-size-fits-all solution when it comes to these platforms. Each tool has a unique set of features, strengths, weaknesses—and price points. LeanConvert, as tool-agnostic optimisation specialists, can help navigate the tricky procurement process and find the right tool or set of tools for your specific business. There’s really no other company in the field that can say the same.

It’s all well and good being data-driven in your decisions by A/B testing, but if the tests themselves are not founded in data, you are not giving yourself the best chance of success. Typically, web analytics platforms are used to ask quantitative questions such as where visitors are struggling, or which areas are in most need of optimisation. Pairing this with experiential analytics tools takes you even a step further in understanding your customers.

Whereas traditional web analytics tools primarily give you information about your customer’s end destination—did they convert? How long did it take them to convert? How much did they spend?—experiential analytics provide much more context by focusing on other actions the user took while on your site. Perhaps they visited the FAQ page, or highlighted the product name before leaving your site, signifying that they copy and pasted it into their browser to try to find a better price. Ultimately, these tools bridge the gap between how you planned your CX and how your customers are actually interacting with your site.

Perhaps even more important than the tools are the people using them. Who are your optimisation champions? These could be new people you’re hiring, or perhaps talent that already exists within your company. Regardless, starting a new optimisation program means starting a new testing culture, and culture is all about the people. It’s absolutely imperative that you have senior stakeholders that actually believe in an experimentation program. And this doesn’t just mean executives who will do lip service and go through the motions; we mean stakeholders that are invested in the power of an optimisation program, fueling that company-wide culture that’s so necessary for success in this area.

The most successful experimentation programs in the world are the ones that aren’t just concerned about immediate ROI. They are prioritising testing, learning, and data-driven development as a way of learning what their customers do or do not want. It’s a way of evolving your customer experience in the long run, not just the short-term.

There is a minimum of four skill sets your team needs to have in order to run a successful optimisation program:

Strategic: Someone who can define long-term goals. How should the roadmap be structured? How does the testing roadmap fit with the company’s web development roadmap? What are your competitors doing with optimisation, and why?

Example roles: Optimisation Manager/Specialist, Product Owner, UX Manager

Analytical: Someone who can determine why you are seeing certain test results and decipher what that means about your customers. Where are the biggest opportunities for improvement? Why do certain journeys or pages work and others don’t?

Example role: Web Analyst

Technical: Someone who effectively acts as the engineer for your optimisation program. How do you build tests in the front end? How do you satisfy technical needs, like building new journeys, changing functionality, etc.? How do you connect your customer data to on-site personalisation?

Example role: Front End Developer

Design: Someone who understands the visuals of a good CX journey. How can you design things to drive the biggest impact? How do you fit what you’re testing with the current look or feel of the brand guidelines for a site?

Example role: UX/UI Designer

Scalability

Having all four of the aforementioned skill sets can be hard to find. On a small scale, generalists can do this job, albeit usually with limitations. A full-time analyst will have a stronger analytical skill set than someone devoting only 25% of their time to analysis. On a larger scale, it makes sense to have an entire team. At some point, depending on the scale of your business and the volume of traffic available for testing on, there may well be enough demand for four or more full-time team members specialising in these skill sets.

You also need to think realistically about whether you have enough capacity in the business to allocate efforts towards optimisation—bearing in mind that doing optimisation right takes time—time that must be taken away from the other activities that existing team members are performing. If the answer is no, then you should outsource these skills to specialised third parties like LeanConvert. This is the most effective way to scale optimisation in the early stages, whether you want to have our managed service or go in the direction of coaching your internal team.

Who Are Your Main Audiences?

First, let’s take a look at your internal stakeholders. Generally, the more included you make them early on in concept development, the more receptive they will be to the learnings and results as they come in. These stakeholders are also a goldmine of new ideas and helpful feedback, which should be welcomed and sought out frequently. And as mentioned previously, you need senior team members who are actually invested in and excited about optimisation.

Especially when first starting, a good way to get people onside quickly is to find specific problems or pain points that key stakeholders have, for example not being able to get something specific built, or being unable to convince the business that a certain change should be a high priority. Fix these problems by building something in the testing tool and proving the value of the desired change (or lack thereof, depending on the test results!).

As for the external audience or end-users, you’ll need to determine pretty early on the audiences you need to tap into to achieve your goals. The best tests will address the needs of a specific customer type. Trying to please everyone is all well and good in day-to-day life, but you should not waste time optimising your CX for customers who barely visit your site.

Customer types could be as broad and straightforward as “New Customers” (i.e. a set of goals around first sign up or first purchase), or as complex as specific personas (e.g. female customers who are interested in a specific product category). The point is to understand what your valuable customer segments need in order to perform a valuable action on your site in order to change their behaviour for the better.

In Conclusion

When it comes to starting a testing program for your company, the things you need to consider are; when (the sooner the better), how (depending on your resources), why (in terms of overarching goals), and who for (looking at both internal and external audiences). On top of that, you also need a plan for scalability.

This may feel like a lot to consider—and it is!—but really, the most important thing is to start, and to do it smartly. Start asking the questions above, find the path of least resistance to getting your first test launched, and do not be afraid to reach out for expert support.

Taking the first step is the hardest part, and at LeanConvert we have years of experience helping businesses go from little or no testing activity to wide, multi-brand, multi-national optimisation programs. From guidance on choosing the right tools and personnel, to creating a winning roadmap, to designing and building the first campaign or set of campaigns—there is bound to be a way we can help.

For more information, contact us here.

And welcome to the world of web optimisation!