When it comes to the growth of an eCommerce, there are many aspects to consider, there is never a single variable that allows you to increase sales by 265% YoY as in the case of JNPR.

JNPR is a French brand of non-alcoholic spirits that aims to transform the way of making an aperitif, giving the opportunity to experience moments of conviviality without sacrificing taste, regardless of the reason why a person may decide to try an alcohol-free and sugar-free alternative.

So let's start by analyzing the areas we focused on to improve some important metrics for the brand:

• sales increase by 265% YoY
• 50% increase in conversion rate
• 4.1% decrease in Acquisition Cost

We started with a structured work of CRO (Conversion Rate Optimization) focused on optimizing the purchase path from user to customer, considering all the intermediate steps through a technical, qualitative and heuristic analysis.

‍Conversion Rate Optimization

Conversion optimization is a systematic way to get those who view the site to do a desired action, such as making a purchase, and consists mainly of 3 macro phases:

1. Research and analysis (quantitative and qualitative)
2. Assumptions and Prioritization
3. Implementation, Measurement, Iteration

Let's briefly review them one by one, in order to share with you some useful tools that you could replicate, once adapted, for your business.

1. Research and analysis (quantitative and qualitative)

The first phase of research and analysis makes it possible to identify exactly:

• Where are the most important problems that hinder growth
• What, in detail, are these problems
• Why are they such

By asking the right questions about the behavior and needs of potential customers, measuring everything they do on site, it will be possible to arrive at the best solutions.

Remember that data is useful only if it leads to insights that can be transformed into concrete test hypotheses.

If we have too much unorganized data without a clear objective, the risk is of achieving the opposite, that is, of being overwhelmed by an enormous amount of data that we do not know what to do with it, risking arriving at a completely wrong solution.

This article will help you avoid just that 🙂

Soon let's see the main steps we followed in the case of JNPR and how on-site optimization has led to:

• an increase in the conversion rate
• an increase in the average order value (AOV)
• and a consequent decrease in the acquisition cost, to which the work of optimizing and scaling the Meta campaigns has also greatly contributed, but let's proceed in order.

‍

2. Assumptions and Prioritization based on impact and cost

The data collected helps to answer the questions in the point above and the insights will lead to the formulation of hypotheses.

Once we have formulated the hypotheses starting from the data collected and prioritized based on the real impact they may have, we begin to perform smart tests and finally obtain the desired results.

3. Implementation, Measurement, Iteration

At Naniza we have developed a framework to implement the above points that can be used for any optimization project, independent of the sector (eCommerce, lead generation, SaaS).

The process to follow to obtain greater conversions is also very similar for different niches and markets.

There are 3 main phases of data collection and analysis, followed by the creation of a complete checklist including all the problems encountered, which we then transform into action points that are prioritized based on impact and cost.

Let's look at the process in detail together.

‍

1. Research and analysis

Step 1.1 Technical Analysis

An essential requirement is that the analytics tools are set up correctly in order to complete a precise data analysis.

Before evaluating other aspects, it is essential that everything is perfectly optimized on the technical side. So-called bugs are the main conversion killer, minimum effort and maximum yield.

Some of the elements from which to start for technical analysis, as we did in the case of JNPR, are:

• conversions for Browser and Device with relative conversion rate
• site performance, including speed analysis
  

This sample report shows users, transactions, and conversion rates by Browser:

‍

What you must observe first of all are the high-level trends. Does a particular browser have a much lower conversion than all the others? If so, you can click on that browser to drill down on the versions and see if there are specific versions that lower the aggregate conversion rate.

To find this data on Analytics, just go to Public > Technology > Browser and OS. You can choose to view data by browser, operating system, screen resolution, and so on.

As far as devices are concerned, the procedure is very similar: Audience > Mobile > Devices

Another important aspect that is one of the first causes of abandonment in the purchase process concerns the loading speed of the site and related contents.

In the case of JNPR, this first technical analysis made it possible to immediately solve some problems that, if neglected, would have made subsequent steps more complex, risking intervening with changes that would actually have been secondary in the light of what was analyzed. Here's the before and after:

‍

Helpful Tips: You can check the load time of your site using tools such as Google PageSpeed Insights and GTmetrix that provide reports on the site's performance and loading rate, indicating bottlenecks. You will then know which areas need attention.

‍Phase 1.2 Heuristic Analysis

This is the closest thing to using opinions to optimize and if done precisely it can lead to surprising results.

We are therefore optimizers who examine a website in a structured way, page by page, based on our experience and what we have seen working together with the best practices known to us.

What we identify or discover through heuristic analysis is not the absolute truth, since it is precisely an informed opinion.

The result is what I call 'areas of interest'. In the subsequent phases of conversion research - qualitative and quantitative research - we try to validate or invalidate the results.

What does the structured analysis of a website look like?

We evaluate each page based on a number of criteria:

• Relevance: Does the page meet user expectations, both in terms of content and design? How can it match what they want even more?
  
  
• Clarity and Immediacy: Are the content and the offer of this page clear? How can we make them even more immediate?
  
  
• Value: is it communicating value to the user? Can we increase user motivation?
  
  
• Friction: What on this page can cause doubts, hesitations and uncertainties? What makes the process difficult? How can we simplify?
  
  
• Distraction: What's on the page that doesn't help the user take action? Is there anything that unnecessarily attracts attention?

In this phase it is also useful to include an analysis of the behavior of users on site by consulting heatmaps, scroll maps, click maps and recording of the user session.

This makes it possible to enrich the quantitative data of the next point by analyzing the behavior of users on the site and visualizing exactly the interactions they had. For example, we can analyze the scrolling depth of the page and identify which contents users dwell on and which they ignore.

This is very useful if a bottleneck has been identified at a specific stage of the purchasing process.

Helpful Tips: Some of the tools that allow you to do this type of analysis are Clarity, Hotjar, VWO.

Very important: as with AB tests, you need to have a sufficient sample of data before you can trust the results. A rough estimate is at least 2000-3000 pageviews. If the heatmap data is based on a few hundred users, the data does not have the same relevance.

Phase 1.3 Qualitative Analysis and Surveys

An aspect that is too often underestimated is the quality of answers and valuable insights that can be obtained from your customers. Why is it useful?

Because it allows you to identify:

• possible doubts that you have not yet identified and that can also be shared by users who are visiting your site now but are blocked in the purchase process
  
  
• differentiating aspects that led them to choose that product and how this allowed them to achieve a change, you could discover new secondary audience segments whose LTV could prove to be higher than the average
  
  
• the aspects that they appreciate the most, even here there are often interesting surprises to be exploited even during the testing phase in the acquisition
  
  
• possible new competitors, even indirect ones, which alternatives have they considered before buying
  
  
• New suggestions to take advantage of for the development of new products

Some of the tools you could use are Google forms (free) and Typeform.

Helpful Tips:

- if the data collection is written (e.g. survey) don't underestimate the power of open-ended questions, listen and pay attention to the words your customers use to describe your product.

- offer a reward to motivate them to respond to the survey, valuing the time they dedicate to you.

In the case of JNPR, thanks to customer interviews and post-purchase surveys, we have collected valuable information that goes beyond quantitative data that can be analyzed by the behavior of users on site, namely:

• how they used the product (e.g. cocktails that we wouldn't have paid attention to if it hadn't been for them)
  
  
• needs to have many different inspirations to create alternative aperitifs, which then translated into repurchases increasing the number of returning customers
  
  
• identify new segments, as in the case of companions who buy the product for their expectant wife, so the target is not only mums-to-be but also and above all dads-to-be.

‍2. Assumptions and Prioritization

The data obtained from the above analyses are the starting point for formulating hypotheses to be translated into tests and optimizations to be implemented in relation to impact, cost and priority.

In the case of JNPR, we created a checklist assigning values to each task in order to be clear which work cycles would be prioritized, based on the impact they could have had. Here's an example:

‍

To formulate the hypotheses precisely, make sure they meet this formula:

We believe that doing [A] for people [B] will make outcome [C] happen. We'll know this when we see data [D] and feedback [E].

Helpful Tips:

Sometimes it is necessary to intervene with changes in the structure and contents of the page, to speed up the process it may be useful to consider using page builders that reduce costs and implement these interventions in the shortest possible time.

Some page builders you might consider for this type of business are Replo and Gempages.

‍3. Implementation, Measurement, Iteration

Once we have prioritized the assumptions and reorganized the work cycles, as in the table above, we can proceed with the implementation.

The mistake that is often made is not to carefully measure the return from what has been done, nor to proceed with the iteration phase based on the output obtained, adapting where necessary.

The work done in these phases has made it possible, in the case of JNPR, to achieve a first major improvement in terms of results:

• sales increase by 265% YoY
• 50% increase in conversion rate
• 4.1% decrease in Acquisition Cost
  

The points above are the starting point for subsequent implementations. Once this is done, we can proceed with the next step: increasing the average order value.

‍How to increase the average order value?

As we said at the beginning of the article, when it comes to the growth of eCommerce, there are many aspects to consider and the variables at play are varied.

After increasing CR by 50% and managing to decrease CAC by 4.1%, another objective for JNPR, as indeed for many eCommerce businesses, it remained to be able to increase the average order value (AOV) and obtain an even greater margin.

Why intervene to increase precisely the average order value?

Because it allows us to:

• incur higher acquisition costs, especially at important times of the year when competition increases
  
  
• improve cash flows by keeping it stable or even increasing margins and profit

But let's take a small step back...

The mistake that could be made in this case is to look at the AOV as a single metric without taking into account the Modal, that is, the distribution of orders and the value that appears most frequently:

‍

From the graph above, it is possible to see how considering the distribution of orders is quite different from taking into account only the average store order calculated as a whole, equal in this case to about 62€.

In fact, the larger order volumes in the example above have AOV of 30€ and 80€ respectively.

Based on this evidence, we have therefore started to develop ad hoc offers aimed at increasing those specific values, such as:

• incentives within the cart to obtain free shipping when a certain threshold has been reached
  
  
• new cross-sells with specific pricing points calculated to reach the target AOV
  
  
• new bundles for customer segments, with the aim of encouraging the purchase of more products
  

all this taking into account which are the customer clusters of the orders in the graph above, for example, are those with AOV equal to 30€ first time buyers or return customers? This affects the type of offer that is going to be developed.

Integrating these and other targeted changes after data analysis made it possible to increase the AOV by 96% in a first phase, and by a further 5% in a subsequent phase.

If you want to learn more about techniques and examples to increase the AOV read this article: Average Order Value (AOV): what it is and how to increase it to make your eCommerce grow

Optimization of Meta campaigns: testing and scaling

There is another fundamental aspect that has contributed to the exponential growth of the brand and the decrease in the cost of acquisition, and that is the work of optimizing and scaling the Meta campaigns.

To be able to maximize the testing and scaling phase of advertising campaigns by optimizing the creation of new content and the recirculation of creativity to be integrated by iterating quickly thanks to high volumes and the amount of data collected with the analyses mentioned at the beginning.

However, this is another chapter that deserves an in-depth study in its own right, so we have written an article entirely dedicated to the testing and scaling part of Meta campaigns where you will find many examples and ideas applicable, once adapted, also to your business.

Click here to read the article: How to create winning Ads and make the creative process scalable.

‍

I hope that some of the points seen so far may be useful to you, if you want to know how to apply them effectively to your business, contact us through the form below 👇🏼