Although Quora existed for few years now, recently I started using it more regularly. Perhaps opening Invesp’s office in Istanbul, Turkey was a major factor there.
And yes, I am a big fan!
Spending my time between our US and Istanbul offices is an amazing adventure. At the same time, it comes with its own set of challenges. That is where Quora comes in handy. I usually ask questions trying to understand more about Turkish culture, economy, business practices and yes, even good restaurants to eat at.
The website has also a healthy section on conversion optimization where those who are looking for advice post their questions. As I went through that section, I noticed that some of the questions were repeated again and again.
So, in this blog post I will take the 8 most common questions on conversion optimization I see in Quora and answer them.
1. What Is the Best Strategy for Increasing My Website Conversion Rate?
You see this question asked in many different forms. Those who run an e-commerce website ask about getting more orders. Those who run lead generation website ask about ways to increase the number of leads from their website.
The problem is that many of the answers focus on tactics. Rarely do you find a thread that goes into strategy. That is fine, if you are looking for quick fixes. However, it comes with a problem.
Conversion optimization tactics may or may not work for your website.
There is a place for tactics but without a well through out strategy, you are most likely wasting your time.
Each web page in your website represents an opportunity. If you are working with an unlimited budget, unlimited resources and unlimited time, then you have all the time to test every little detail on your website.
If you are like the rest of us, then you want to start fixing these pages that present the most impact on your bottom line.
How do you find the pages with the highest potential?
To increase your website conversion rates, you should use analytics to understand how visitors are flowing from one area of your website to the next.
In order for you to get to the big “Yes” – a macro conversion – the visitor must go through a series of smaller “Yeses”, or micro conversions. Before visitors place an order with your website, fill out a contact form, or subscribe to your service, they go through a series of steps and follow different paths.
Our task as marketers is to identify all the possible funnels which visitors go through and to determine which of the funnels are working well and which funnels are leaking visitors so that we can focus on them.
The next question to think about is what changes should you make on a particular web page?
I will tell you what you should NOT do first. You do not randomly change things because best practices tell you to do so. Persuading visitors to convert requires us to anticipate the information they need and to present that information to visitors when they need it.
So, how do you identify conversion problems on a webpage?
At Invesp we use the Conversion Framework to determine the factors that impact visitor behavior on a web page. These factors are grouped into two main categories:
- Website-centric factors
- Visitor-centric factors
Website centric factors are the factors that relate to your website and the signals it sends to your visitors. The main categories under website-centric factors are:
- Trust factors: Remember that if a visitor does not trust you, he will not interact with you.
- FUDs: How do you identify and minimize the fears, uncertainty, and doubts that stop website visitors from doing business with you?
- Incentives: How do you incentivize a visitor to interact with you immediately?
- Engagement: How do you engage visitors so you move from a one-time interaction to a long term relationship?
Visitor-centric factors are the factors that relate to your website visitors, the nature of their purchase and where they are in the buying funnel. You have less control over these factors, however, you must account and prepare for them.
The main categories under visitor centric-factors are:
- The visitor persona: How do you address the different types of visitors’ temperaments?
- How do you design your website to address each of the different visitor buying stage?
- The third factor is the complexity of the sale. Selling an item that costs $10 is a lot different than selling an item that costs $2,000
We typically evaluate every page (or class of pages) on your website using the Conversion Framework with the goal of determining possible conversion problems on that page.
A conversion rate optimization specialist can identify 50 to 150 problems on a single webpage.
You then prioritize these problems and focus on the top 3 to 5 problems by creating new designs to fix them.
But you do not stop there. You conduct A/B or multivariate testing by splitting your current website visitors between your original design and the new designs and watch which of these design generates more conversions.
Quora question: What’s the best ways to increase your conversion rate honestly?
2. Tools to Perform a Conversion Optimization Audit of My Site?
As of now, there are no tools that help you conduct a conversion optimization audit of a website. There are, however, tools that help you understand how visitors behave on your website which, in turn, will help you in the process of conducting a CRO audit for your website.
Some might think that I am splitting hairs. But a tool that is designed to conduct an audit will be designed in a completely different way compared to a tool that tracks how visitors behave on the website.
Having said that, here is a list of tool categories which we think every website owner should have installed on their website:
Analytics: Needless to say that analytics is a must have for every website. Any self-respecting analytics tool will help you track the performance of different pages on your website (how many visitors they are getting, what is the bounce and exit rates, etc). The main dominant player in this field is Google Analytics. And while GA is free (there is an enterprise option), you are giving your data to Google.
Suggested analytics tools:
Heatmap: Heatmap software record the clicks visitors make on any page of your website. They also provide a scroll map which shows how much visitors scroll when they view different pages of the website.
Suggested heatmap tools:
Video recording: Similar to heatmap software, the tools in this category record the interaction between website visitors and the different pages on the website.
Suggested video recording tools:
Website survey: the tools in this category ask website visitors to answer a small set of questions with the goal of gaining insights from these visitors on what is working and what is not in your website design.
Website survey tools:
A/B testing tools: these tools allow you to conduct A/B and/or multivariate testing on your website so you can determine which design generates more results for you. We have written a lengthy article on selecting the best A/B testing tool for your website. You can find it here.
A/B testing tools include:
Quora question: Which tools do you use for conversion rate optimization?
3. Does the CRO Pay Per Performance Model Work?
We have not seen it work in the long run and we have been doing conversion optimization for over 11 years.
The idea of a pay per performance tests sounds good at the surface. The client will only pay for increases in conversion rate which shifts the burden and the risk of the project to the CRO firm. For the CRO firms there are two main reasons this model is appealing. First, if the project is successful, then the reward for success will be higher than charging a fixed fee. Second, this model works well for companies that are just starting out.
Here is the problem with this model. The success of any conversion optimization project depends on many factors. But the main two factors are:
- How many tests do you deploy in a given time period and,
- How sophisticated are these tests?
Most CRO firms end up relying on the client’s development team to implement the test. And here the problem lays in. The CRO firm is responsible and will be rewarded for the success of the project, but it will have to rely on the client to implement the tests.
In order for the CRO firm to deploy more tests, the firm must have access to the client’s website and should be able to make changes to it. The need for the backend of the website and its database becomes more prevalent, as the tests get more sophisticated.
But most clients, if not all, will not give unlimited access to the CRO firm because that assumes two things:
- The CRO firm has good handle on the technology used by the client and,
- The CRO firm has the technical knowledge to be trusted with making change on the website.
For any medium to large website both conditions are typically a no go.
So, most CRO firms end up relying on the client’s development team to implement the test.
And here the problem lays in. The CRO firm is responsible and will be rewarded for the success of the project, but it will have to rely on the client to implement the tests.
Some might say that the client is incentivized to see the project succeed and is motivated to implement tests faster. In our experience, that might have been the case but as the project progresses, development teams get busy and CRO work is pushed further behind and is given less priority.
As a result, these projects have an inherited conflict of interest built into them.
We believe that in order for a project to succeed, everyone must have skin in the game. For the client that will include paying a fee.
Now, there are instances where we agreed to lower our fees in return for an additional fee which the client pays, if we meet certain performance criteria.
4. Running Multiple Overlapping A/B/n tests Simultaneously on One Page?
As you talk to different companies, you will hear differing opinions on this matter. It has been in our practice that multiple A/B tests on larger websites where 100% of the traffic is included in an A/B test can and will cause data pollution.
While we recommend running multiple tests at the same time, we also highly advice that:
- The tests are not run concurrently, or
- If tests are run concurrently, each test runs in a separate swim lane.
If tests run in separate swim lanes, traffic from one test does intersect traffic from the other test. This rarely happens on large complex websites where visitors jump from one swim lane to the next.
Good examples of concurrent tests include: running separate tests on desktop and mobile websites, running concurrent tests targeted for different countries, and running concurrent tests targeted for different sources and media of the traffic.
5. What Does “Confidence Level” Mean in Conversion optimization Lingo?
Statisticians have been debating confidence level for over 90 years.
Confidence level refers to the frequency which an observed interval contains a parameter (a test result). So, if confidence levels are constructed across many separate data analyses of different experiments, the proportion of such intervals that contain the true value of the test result will match the given confidence level.
The term confidence level is one of the least understood, yet most used terms in conversion optimization.
Confidence level does help in persuading people to test, but you need to understand how it works and affects your data.
Regina Nuzzo reports:
According to one widely used calculation, a P value of 0.01 corresponds to a false-alarm probability of at least 11%, depending on the underlying probability that there is a true effect; a P value of 0.05 raises that chance to at least 29%. So Moty’s finding had a greater than one in ten chance of being a false alarm. Likewise, the probability of replicating his original result was not 99%, as most would assume, but something closer to 73% — or only 50%, if he wanted another very significant result. In other words, his inability to replicate the result was about as surprising as if he had called heads on a coin toss and it had come up tails.
In calculating financial risks, the banking sector is sometimes required to see a 99% confidence level. The FDA also requires a 99% confidence level in some cases of drug testing.
Our goal is not to confuse you.
So, what does confidence level mean?
We recommend you think of the confidence level (p-value) as the probability of having a result as opposed to the potential frequency of that result.
So, what confidence level should you choose?
We recommend a 95% confidence level, but if an experiment runs for multiple business cycles and it repeatedly shows higher than 85% confidence level, you should definitely consider it.
Quora question: What does “confidence” mean in conversion optimization lingo?
6. How Do I Decrease the Bounce Rate of My E-Commerce Website?
Let’s start by defining what a bounce rate is. Bounce rate is the percentage of your visitors who land on a web page from an external traffic source and leave the website without visiting a second page.
So, if a web page gets 100 visitors from an external traffic source (google organic traffic for example) and of those 100 visitors, 35 visitors decide to leave the website without ever visiting a second page on the website, then the bounce rate for the page is: 35/100 = 35%.
A common mistake we see with different websites is considering bounce rate for the website as a KPI (key performance indicator). For most websites, it is not.
A KPI should have a direct business impact. Bounce rate and exit rate are important metrics to track on your website. They rarely have an impact on a firm bottom line.
The conversion rate for a website or the average order value, on the other hand, are business KPIs. They impact your bottom line.
All these things make me doubt the general understanding of this question. An e-commerce website should focus on increasing its conversion rates, visits per value or average order value. Optimizing for bounce or exit rates is like chasing a ghost and will have no direct impact on the bottom line.
7. What Do I Need to See During My A/B Testing to Be Sure That One Variation Is the Winner?
There are three different criteria which must be met before you can call a winner for an A/B test.
- You need a statistically significant winner. Statistical significance will provide you with an indication on how solid your test is. The higher the confidence level for the challenger, the less likely it is that the challenger is a false positive. For most tests, a confidence level of 95% is acceptable. There are however instances where the challenger reportedly remains at 80% confidence level or a bit higher. In these cases, you can still call the test with the challenger being a winner, knowing full well that there is a segment of the visitors which will likely convert better with the control (original design).
- Both the control (original design) and the winning challenger (variation) must have recorded a minimum number of conversions: the goal from this criterion is to ensure that your challenger is NOT a false positive and that testing data stabilized. The minimum number of conversions will vary from one website to the next and will depend on the number of conversions your website gets per month. With smaller websites, we usually ask for a minimum of 100 conversions per design. For larger websites, the minimum can be as large as 3,000 conversions.
- The test should have run for at least one week. Website visitors shop differently on different days of the week. At a minimum, you want the control to run against the challengers for at least one day in the week. If you want to be more cautious, you can increase the minimum time required to two weeks.
Another thing we recommend in cases where the test includes many challengers against the control is to run the winner of the test against the control in a head to head test. This new test will help ensure that your winning challenger actually beats the original.
8. What’s a Typical Conversion Rate for Online Retailers?
In the early 2000’s, Shop.org reported the average e-commerce website converted 3% of its traffic into customers. FireClick Index data from 2015 show that an average e-commerce conversion rate is close to 7.2% globally.
|Business Metrics||Last Week|
|Conversion Rate: Global||7.20%|
|Conversion Rate: First Time Visitors||7.60%|
|Conversion Rate: Repeat Visitors||6.90%|
|Cart Abandonment Rate||53.60%|
Conversion data for fashion and apparel, from Fireclick index:
|Business Metrics||Last Week|
|Conversion Rate: Global||5.90%|
|Conversion Rate: First Time Visitors||6.10%|
|Conversion Rate: Repeat Visitors||5.60%|
|Cart Abandonment Rate||N/A|
Conversion data for catalog websites, from Fireclick index:
|Business Metrics||Last Week|
|Conversion Rate: Global||6.70%|
|Conversion Rate: First Time VisitorsC||6.90%|
|Conversion Rate: Repeat Visitors||6.20%|
|Cart Abandonment Rate||55.20%|
Conversion data for specialty websites, from Fireclick index:
|Business Metrics||Last Week|
|Conversion Rate: Global||7.60%|
|Conversion Rate: First Time Visitors||7.30%|
|Conversion Rate: Repeat Visitors||8.80%|
|Cart Abandonment Rate||50.20%|
Conversion data for outdoors and sports websites, from Fireclick index:
|Business Metrics||Last Week|
|Conversion Rate: Global||4.80%|
|Conversion Rate: First Time Visitors||4.90%|
|Conversion Rate: Repeat Visitors||5.10%|
|Cart Abandonment Rate||60.10%|
Conversion data for software websites, from Fireclick index:
|Business Metrics||Last Week|
|Conversion Rate: Global||4.10%|
|Conversion Rate: First Time VisitorsC||4.20%|
|Conversion Rate: Repeat Visitors||3.60%|
|Cart Abandonment Rate||76.10%|
Back in 2010, outdoor and sporting stores converted less than 0.5% of the traffic they receive. In 2015 these stores were converting 4.8% of their visitors into customers. This is close to 900% increase in conversion rates.
The purpose of this data is to help you compare your site to the average conversion rates of your industry. But you shouldn’t necessarily feel better, if you are within your industry average because most online conversion rates are scant compared to brick-and-mortar rates.
Quora question:What’s a typical conversion rate for online retailers?