You feel you have done everything right on your website. The copy looks compelling. The design looks magnificent. But still, visitors are not spending as much time on your website as much as you want them to. You have little to no clue about why it is happening and how to fix it. You try out a bunch of SEO tools, you undergo usability testing, but still, no significant improvements can be seen.

You feel tired and reluctant to try out more SEO tools and do more testing. You still have no clue why visitors are abandoning your well-designed information-rich website. You feel like talking to your visitors to understand what’s wrong, but how is that even possible? They certainly are in no mood to fill out a form on an exit popup or converse with your chatbot before leaving the website. The users that abandoned your website must have experienced issues you were not aware of.

Using a Real User Monitoring Tool

So, what’s next? With the attention span of people decreasing day by day, solving this problem for which you do not even know the root cause is imperative. Here’s how you can understand what’s going on. Try integrating your website with a real user monitoring tool.

Real user monitoring, an approach to web monitoring, is a type of performance monitoring that captures and analyzes every transaction of the user. It is a form of passive web monitoring because the monitoring services run in the background. There’s another type of web monitoring known as synthetic monitoring. In synthetic monitoring, the monitoring scripts are deployed via the browser to simulate paths that website users take.

Real user monitoring technology looks at how visitors interact with a website or an application. Real user monitoring tools enhance the front-end performance of your website by analyzing data to detect anomalies from real user sessions thereby increasing customer satisfaction. These tools can also send you alerts in real-time. Using these tools, you can get to know the exact underlying issues concerning your website. 

A real user monitoring tool like Sematext Experience is easy to use. It alerts you when user experience gets affected by website performance. You can inspect page-level specifics, track page loads, HTTP requests, UI interactions, resources, and more. Using such a tool, you can reduce performance-related issues by having 100% visibility into what areas are affecting UX.

How Real User Monitoring Improves SEO?

Real user monitoring lets you know the problematic areas on your website which could be affecting user performance. Here’s how such a tool can help in improving SEO.

Improving Page Loading Time

Google indicates that the loading time of web pages is a signal used by its algorithm to rank them. Google could be measuring the time taken to receive the first byte while calculating page speed. A poorly performing website degrades the user experience. Hence, such sites do not deserve to be promoted in search results.

Time-to-first-byte (TTFB) is the time taken by the browser to receive the first byte from a web server when a URL is requested. This metric includes the time taken to send a response, the processing time of the server, and the time taken by the first byte to reach your browser from the server. Here’s a graph of median TTFB vs ranking positions:

Source – https://moz.com/blog/how-website-speed-actually-impacts-search-ranking

Sematext Experience collects resource timing data for images, CSS, and JavaScript files. When someone visits your website, the browser begins downloading resources. These resources have a big impact on your website’s loading speed.

The resources are shown in the form of a waterfall ordered by the time they were discovered.

Resource Waterfall – Sematext Experience 

Load-time performance depends on various factors. If visitors experience slow page-loads or poor HTTP request performance, it is most likely caused by network latency, DNS servers, redirects, application performance, and other factors. A detailed timing breakdown on page loads and HTTP requests is displayed to understand which part(s) of the loading process is slow. 

Timing Breakdown – Sematext Performance

We can only optimize what we can measure. Once you have information on the resources which are slow-loading, you can start improving their load times. As SEO is impacted by loading time, here’s a list of tasks you can do to optimize your web pages for better SEO:

• Reduce the dimensions of images to fit the image containers on your web pages.
• Compress images to further reduce the space they occupy on your server.
• Stick to standard image formats (JPG, PNG, GIF).
• Use CSS sprites for images that you use frequently on your website like icons and buttons. A CSS sprite loads entirely in one go thereby reducing the number of HTTP requests. This technique helps in saving load time by not making users wait for multiple images to load.
• Reduce the size of your HTML, CSS, and Javascript files using a minification tool.
• Remove code comments, formatting, and unused code to reduce the size of the files.
• Reduce the number of redirects to avoid additional waiting time.
• Move scripts to the Footer to avoid render-blocking javascript code.
• Use browser caching so the browser doesn’t have to reload the entire page when a visitor comes back to your site. Browsers can cache images, stylesheets, Javascript files, and more. Learn how to leverage browser caching here.
• Improve the server response time by looking out for performance bottlenecks like slow database queries and routing. Try switching to a new hosting solution to further improve the server response time. The optimal response time is under 200 ms.
• Use a content distribution network (CDN) which essentially distributes the load of delivering content. It works as follows – copies of your website are stored at multiple data centers located geographically so that users have faster and reliable access to your website.

Improving User Satisfaction

The more the users are satisfied with your website experience, the more likely they will spend a lot more time on your website which in turn will help in improving SEO as time spent on the website directly affects your search engine rankings. Naturally, we want the average time spent on your website to be as high as possible for Google to take note of the importance of your website. 

On the other hand, if users are dissatisfied with your website experience, they are more likely to abandon your website which eventually will lead to an increase in the bounce rate. Bounce rate is the percentage of visitors who abandon your website after visiting a single page. A high bounce rate means that visitors don’t like what they see on your website. We want to have the bounce rate as low as possible to have better rankings.

You learned how time spent on the website and the bounce rate impacts the rankings for your website. A real user monitoring tool like Sematext gives a measure of user satisfaction. 

User Experience – Sematext Experience

The user satisfaction score is based on the Apdex industry standard. Page loads, on-page transactions, and ajax requests are separated into different user satisfaction scores as they have different performance characteristics.

Conclusion

As 68% of online experiences begin with a search engine and 53.3% of all website traffic comes from organic search (source), it is important to improve the SEO of your website for better rankings. The fact is that there could be issues concerning your website which you might not even be aware of. These issues could secretly be affecting the user experience thereby making visitors abandon your website. You need real user monitoring tools to know these issues to improve your website rankings. Do research regarding which real user monitoring tool is good for you – one that fits your use case. The important part is to monitor the end-user experience and be in control of how users interact with your website.

You probably heard someone talk about serverless technology around the office or at some conference conference your boss made you go and while understanding it’s merit as a technology doens’t come easy for a person without a technical background. With that said, I’d like to go over the biggest benefits of moving away from a traditional server and stepping into the cloud.

At its core, serverless is not the lack of servers but the existence of a server that does not need to be managed. Using microservices uploaded to Lambda that executes a little bit of code at a time and returning a value, you can build large-scale enterprise level application without having to see a single line of PHP. Now, this might not make everyone as happy as it does me but for a front-end developer like myself, that’s a godsend.

So why is this such a big deal?

It stands to reason that any new technology trying to challenge the old ways of creating, running and hosting web and mobile apps will have a tough time getting support, yet the serverless community has grown exponentially in the past year and the reasons behind the movement are clear.

• Lower costs
• Faster deployments
• Graceful scaling

How does this affect me and my store?

Well, there are a number of major upgrades that a serverless approach can bring to your store. Let’s start with the obvious. The cut in costs. Using a serverless technology is cheaper than your traditional hosting and if you ever hosted a Magento store (or basically any other self-hosted solution) you’ll know just how expensive running, even a semi-successful store can be.

For AWS Lambda, the monthly compute price is $0.00001667 per GB-s and the free tier provides 400,000 GB-s. The monthly request price is $0.20 per 1 million requests and the free tier provides 1M requests per month.

There are more and more companies sharing their stories on how the cut costs using serverless and if you want to you can read about how companies save money with serverless.

The second reason why going serverless can be of significant value is the unparalleled scaling that can be achieved with serverless. If you do any online shopping around the holidays, especially during the “Black Friday” sales you probably saw your fair share of stores that were down, crumbling under the weight of countless users trying to browse, add to cart or compare products.

Creating microservices for all your websites functionality will improve your page load speed, remove scalability concerns and making changes to your code or business logic will be far easier to implement.

How would this even work?

So how exactly would you go about switching your store from a traditional server and going serverless?
To make things easier for everyone to understand I’ll won’t get too technical as doing so will turn this article into a 15K words software documentation, and let’s face it. I’m not smart enough to pull that sort of thing off. What I will do is break the store into small sections and explain how I would tackle the situation.

1. The CMS

I’m referring to the static pages like homepage, contact page, about us, etc. For this section use a caching system to create static pages from the CMS and have them served from a cloud service provider like AWS or Microsoft Azure. Most e-commerce content management systems (CMS) already come with a half decent caching solution so you might already be familiar with the concept.

Personally, I’d go with Amazon EslastiCache. It’s easy to install and deploy and won’t break the bank.

Amazon ElastiCache offers fully managed Redis and Memcached. Seamlessly deploy, operate, and scale popular open source compatible in-memory data stores. Build data-intensive apps or improve the performance of your existing apps by retrieving data from high throughput and low latency in-memory data stores.
— Amazon ElasiCache

To keep this article short and sweat I’ll link[http://blog.e-zest.com/how-to-implement-aws-elasticache-in-magento/] to a tutorial on how you can integrate ElastiCache with Magento(as this is one of those platforms that just won’t work without proper caching).

2. The catalog

The catalog is the part with all the categories and products and the main issue you have with them is the rendering speed. Having thousands of products with filters applied will slow your website to a halt. Now imagine having 10, 100 or 1000 concurrent users browsing, filtering and comparing products.

The solution? Microservices!

Here is where AWS Lambda shines brightest. Every interaction that the user can have with the website is split into small functions. So having 10 people on your site or 1000 won’t make a dent in the page load time as AWS will scale to best fit your needs.

But wait, there’s more!

The best part of this approach is the fact that you only pay what you use. That’s right. No more paying for big server bills when your store gets no visitors. Here’s a neat little AWS pricing calculator to help you figure out what to expect in terms of cost.

Static files, like images and videos, can be served from S3 or even better: CloudFront to reduce costs and increase speed.

3. The cart

This is probably the subject that most people that stayed this far along in the article are interested in. Carts are extremely complicated and cumbersome systems that will eat that CPU as if it were made out of tacos. The problem with conventional stores is MySQL bottlenecks and when that happens you entire website stops responding until MySQL catches up(either that or the entire thing just goes down).

The fix is not that complicated. Stop using MySQL! Since you are already using AWS Lambda why not use DynamoDB, it’s easy to setup, cheap and reliable and before you start calling me a sellout for pushing AWS on you all, MongoDB is another solid option, in fact, I kind of prefer it to DynamoDB as I’m more used to using it. Here’s a quick guide on how to get started with Lambda and MongoDB.
Besides moving away from using MySQL, I’d urge anyone that gets a lot of orders to consider using an API like Magento’s API to handle orders. At the moment, having an expensive server with all the bells and whistle can withstand 3 to 5 orders/second which as it turns out, it’s still less than stores like Angry Birds store or Anki.com need.

The way it would work is as simple as it is efficient. You handle the order queue via a javascript based checkout system that interacts with the store API validating and recording each order. Here’s a tutorial on how to do this in Magento: https://devdocs.magento.com/guides/v2.1/get-started/order-tutorial/order-intro.html

So what’s the problem?

Why isn’t anyone and everyone using this solution yet? There are a lot of people that already use this solution and it saves both time and a lot of money. So why are there still people not using this? One reason might be the fact that serverless is still a new technology. Not everyone feels comfortable with it yet and because it’s so new the tools needed for development are not popular enough. Take for example AWS Lambda debugging. It’s a basic need for every developer and yet Amazon doesn’t have a tool that provides all the support needed for development. So we end up looking for tools Like Thundra.io and Dashbird.io(amongst others).

In conclusion

Moving from a traditional self-hosted solution to the cloud will be a complex problem for store owners but at one point or another, they will find themselves having to face difficult decisions vis-a-vis their platform and scaling capabilities and going serverless can solve many problems.

I’m anxious to see what’s to come in the e-commerce space. Perhaps we’ll see our first in-the-cloud e-commerce solution. The sky’s the limit! (get it? cloud computing – sky? ok, I’ll stop).

A couple of weeks ago I was invited to an investment dinner where 3 startups made their pitch in front of a panel of investors. The event turned out great and the startups that participated were well prepared to answer all of our questions, technical or otherwise. One thing bothered me, perhaps more than it should. Two out of the three startups boasted about their “amazing” AI even though they were merely a run of the mill machine learning script that was interpreting a set of data based on a predefined algorithm.

Most people don’t understand the difference between the two, some realize there is a difference but can’t really point it out and there are those that understand the relationship between Artificial Intelligence (AI), machine learning (ML) and Deep Learning(DL – to be discussed at a later date).

To put things simply, machine learning is a technique used to achieve artificial intelligence. Arthur Samuel coined the AI term and defined it as “the ability to learn without being explicitly programmed.” Without Machine learning, it would be a lot harder to create an AI, you’d have to write complex rules and decision trees that would take a lot of programming so instead, we use machine learning to train our AI friend to learn on its own by feeding him large amounts of data.

On the other hand, machine learning is an algorithm used to process large amounts of data and act on the result. Deep learning is one of many ways of approaching machine learning, but we’ll get to that at a later date.

To say artificial intelligence equals machine learning is just juvenile, plain and simple. True, ML is a component of AI but learning the difference is important.

Let me paint a scenario that would make more sense. Let’s imagine going to the market to buy your favorite apples. You get to the table where the lady usually sells them to you and you ask for for a couple of apples but in return, you get two apple seeds. She then proceeds to explain that you can use the apple seeds to eventually get apples but technically it’s not the same thing. Right? The same principle applies to machine learning vis-a-vis to artificial intelligence. ML is a means of achieving AI, not the other way around. AI can exist without ML by writing an immense code that would guide the computer into making decisions by itself.

In this video, Reza Zadah clarifies the difference between artificial intelligence and machine learning, and what role algorithms play in these fields.

I strongly believe AI will paint our future and while technology is not quite there yet to make ai available to everyone, it will be in a few short years and it’s important to educate ourselves and learn the possibilities it can provide and explore the ways it cna improve our lives.

For the past few years, I’ve seen a lot of companies talking a big game when it comes to machine learning and artificial intelligence and I think it’s imperative to understand the difference between the two so we can stop wasting time debating articles like this one and focus on creating solutions that will benefit us all. Thankfully we have companies like www.typingdna.com,  www.groupbyinc.com, and www.baidu.com that push the envelope further with amazing solutions that promise to change our daily routines for the better.

2018 is right around the corner and I thought I’d start celebrating the great year 2017 has been a list of the most promising startups in tech that you should look for in the coming year. And before I start dropping names left and right, I do want to point out that this list is purely biased and strictly based on my uneducated opinion. So without further ado here are the top 7 tech startups I’m looking forward to seeing in 2018.

1. Zentoshop

Zentoshop is a Magento based e-commerce Saas that allows store owners to go online in a matter of minutes and the guys behind the product have managed to solve one of the biggest issues I had with Saas E-commerce platforms and that would be scalability. Normally, a store you build on a similar platform would work (probably great) for the first few months, maybe a year, but the minute you hit big traffic numbers, orders or have a lot of products, the platform would crumble under its own weight. Zentoshop having Magento under the hood solves this issue, allowing the user to take advantage of all the benefits of having a Magento shop without any of the drawbacks, which are mostly the technical issues that force Magento store owners to hire one or more developers to do regular maintenance and updates on the system to ensure everything’s running smoothly.