ImageEngine vs Cloudimage — Which is the More Powerful Image Optimizer?
This is part II of this comparison series, you can find part I here where we look at the user experience and features.
On the internet, speed means everything. It’s the difference between a frustrating user moving to the next search result — or even ranking high on Google SERPs in the first place. However, as our online presence grows and we use more media on our pages, our web pages are slowing down.
It’s no wonder that tools like ImageEngine and Cloudimage are becoming increasingly popular to accelerate web pages, without having to get rid of your precious images.
In the second part of this two-part series, I’m putting both these image optimization services to the test. We’ll look at their potential to reduce the performance impact of your images, how they achieve it, as well as how they stack up in terms of value-for-money.
Ready, set, go!
Performance Test — What Results Can You Expect From Cloudimage and ImageEngine?
Considering that the main reason for using a service like Cloudimage and ImageEngine is to improve your website performance, it’s logical that you’d expect to see a tangible return on investment for what you pay.
So, to see what kind of improvements you can expect, I set up a simple test using Google’s PageSpeed Insights. PageSpeed Insights reveals not only useful information like how large your website payloads are and how long it takes to load, but also critical user-centric performance metrics. These “core web vitals,” as they are dubbed by Google, now even play a role in your SEO rankings on top of giving an insight into the user experience of visiting your website.
For this test, I created a basic web page with a large number of high-quality, oversized images without any manual optimization. After running a PageSpeed Insight audit, here were the results:
Despite being a straightforward page with no complicated layout or interactivity, the score is less than optimal. The overall Speed Index and LCP is particularly hard hit by the weight of the images.
Both give an overall impression of how long it takes the page to visually load for users. Three seconds is considered the critical threshold where you’ll start to lose a lot of incoming traffic due to increased bounce rates. Currently, LCP is one of Google’s core web vitals along with CLS (Cumulative Layout Shift) and FID (First Input Delay).
When digging into the diagnostics, we can see how images contribute to this sub-par score:
As you can see, the biggest problem is the enormous network payload thanks to all the images — roughly 61 MB. As Google points out, we can bring this down by:
- Properly resizing images
- Serving them in next-gen formats
- Efficiently encoding images
Cloudimage PageSpeed Insights Test Results
Ok, so now that we have our benchmark and metrics to improve, let’s see how Cloudimage does. For the first test, I only used Cloudimage’s default optimizations to see what it has to offer as an out-of-the-box solution:
As you can see, this didn’t do much and led to some other issues when it comes to the time it takes the page to become interactive. Despite a worse overall score by PageSpeed Insights, Cloudimage did still manage to reduce the image payload by just over 50% to around 30 MB:
Cloudimage did take care of using next-gen formats and efficiently encoding images. However, as you can see, Cloudimage can’t dynamically properly resize images without the aid of responsive syntax.
To give Cloudimage the benefit of the doubt, I installed the JS plugin that implements responsive syntax as well as the lazy-loading module. After a bit of work, the results were far better:
Both the overall PageSpeed Insights score and the individual metrics were greatly improved, particularly the crucial LCP. However, there was still some residual CLS despite the taking Cloudimage’s recommend steps to prevent it.
There was still a bit of room for improvement when it came to properly resizing images. This is probably due to Cloudimage’s reliance on responsive syntax with a limited set of breakpoints:
Finally, Cloudimage managed to reduce the payload to around 2.3 MB, a massive 96.4% reduction.
While these results are superb. There are a number of caveats:
- I had to integrate the responsive syntax plugin, which involves taking even more time and effort to modify your web project files and image tags.
- This solution involves JavaScript, which may introduce other issues down the line to do with browser compatibility, breaking pre-loading, and a JS-related performance hit.
- Technically, you can use a similar solution using responsive syntax and improve the results for any other image CDN as well.
ImageEngine PageSpeed Insights Test Results
For testing ImageEngine, I simply replaced the src attributes of my images with my ImageEngine delivery address. Without any help of responsive syntax, this was the result:
ImageEngine achieved a nearly flawless performance, with great scores across all the metrics. PageSpeed Insights seems to have agreed as there were no more opportunities for improvement listed. This means that ImageEngine took full advantage of properly resizing images on top of efficiently encoding and serving them in next-gen formats.
As the results show, ImageEngine by itself is good enough to match or even exceed another image CDN + responsive syntax.
Because ImageEngine uses intelligent device-detection technology, the image optimization engine itself can dynamically resize images based on the context they are viewed in. As it doesn’t rely on predefined breakpoints, it can adapt to a much wider variety of scenarios with fine-tuned resizing.
As an added bonus, ImageEngine also managed to reduce the total payload even further to around 1.5 MB (97.5%):
Pricing
Both Cloudimage and ImageEngine work based on a similar double-pronged pricing structure. Firstly, you pay an upfront monthly fee that comes with a preallocated amount of CDN bandwidth (and image cache for Cloudimage only). If you go beyond that limit, you’ll be charged according to the overage rate.
Cloudimage’s pricing is divided into five plans that include a Free tier, three paid plans, and an Enterprise package. Startup and Pro users can subscribe annually or on a month-to-month basis, with the first option working out around 10% cheaper. However, Rocket and Enterprise plans are annual-only:
So, for example, if you are subscribed to “Pro” and consume around 200 GB of image traffic, you’re monthly bill would look like this:
$79 for first 150 GB + ($0.5 GB*50 GB overage) = $104/month
It’s important to note that CDN traffic is based on the transformed image traffic served to visitors. So, if an original image was 2 MB in size but was optimized down to 1 MB, you will only consume 1 MB of CDN traffic each time that same optimized version was requested.
Pretty much everything is included with all plans, such as all image transformations, image optimization, and responsive images plugins. Only the Free plan is excluded from premium support access as well as the 99.0% SLA agreement.
One point about Cloudimage that sticks out like a sore thumb is that it limits and charges you for the amount of image cache you take up on their servers. This consists of all original images plus their transformations that are generated and cached by Cloudimage.
This cache will tend to grow over time as more and more variations of the origin images are requested. However, Cloudimage does use a cache-cleaning algorithm to flush inactive images from your image token’s cache (although they remain in the CDN cache in case they’re needed). While this will help to slow down the snowball effect, you’ll still have to track your cache usage and manually invalidate the cache when needed to avoid extra costs.
ImageEngine, on the other hand, has four pricing plans: Free, Basic, Standard, and Pro. The Free plan is available for 30-days and all other plans have this 30-day free plan built-in. All plans are month-to-month, so there are no tie-ins or contracts to worry about. Like Cloudimage, all plans come with maximum image transformation and optimization capabilities.
One of the small differences is that you can create CNAMEs with ImageEngine’s Standard plan while Cloudimage’s Pro plan comes with only 2 and the pricier Rocket plan with 4.
In terms of the pricing rate, the Basic plan starts at $49 for 100 GB of Smart Bytes. Smart Bytes are the CDN traffic consumed by serving optimized images. The Standard plan comes with a huge 250 GB of Smart Bytes for $99/month.
While ImageEngine and Cloudimage’s pricing plans don’t line up in terms of raw cost, you typically get more bandwidth mileage out of ImageEngine at a lower price point. For example:
ImageEngine Free: 100 GB / $0 for 30-days
ImageEngine Basic: 100 GB / $49 ≈ 2 GB of bandwidth per $1 spent
ImageEngine Standard: 250 GB / $99 ≈ 2.5 GB of bandwidth per $1 spent
ImageEngine Pro:
- Pricing scales to meet Smart Byte usage and tailored requirements
- Includes all features of Standard plan plus:
- WAF with DDoS protection available
- Custom reporting available
- Dedicated edge servers and optimization engines available
- Ticketed enterprise support or optional managed service
- Custom service level agreements
Compare that with Cloudimage:
Cloudimage Startup: 40 GB / $29 ≈ 1.4 GB of bandwidth per $1 spent
Cloudimage Pro: 150 GB / $79 ≈ 1.9 GB of bandwidth per $1 spent
Cloudimage Rocket: 800 GB / $249 ≈ 3.2 GB of bandwidth per $1 spent
ImageEngine also has a lower flat rate for bandwidth overage at $0.4 per 1 GB.
The only exception to these rules is the Rocket plan from Cloudimage. However, this plan is disproportionately pricing compared to the others and is really only for extremely high-traffic web apps.
ImageEngine’s superior bandwidth mileage is stretched even further if you take into account the image optimization results above. As ImageEngine reduces payloads by more than Cloudimage, you’ll probably use even less relative CDN bandwidth for the same amount of raw image content.
Another great feature about ImageEngine’s pricing is that it only depends on your bandwidth usage and nothing else. Like Cloudimage, they don’t track the number of transformations you make or the original image count. However, ImageEngine also doesn’t take into account how much cache you take up on their servers.
While the actual cost impact of this is highly situational, it will provide greater peace of mind at the very least and real cost savings at best.
Conclusion
Although there is a lot to unpack here and the best choice will ultimately depend on your individual needs, we can make some general conclusions:
- Cloudimage seems to be better at accommodating customers at the extreme ends of image optimization needs and budgets. The Free plan should be adequate for relatively low-traffic sites without tonnes of high-quality images. On the other hand, the Rocket plan is aimed at enterprise-grade web apps with huge traffic volumes and little-to-no budgetary restrictions.
- For most users, in most situations, ImageEngine will most likely be the most cost-effective image optimization and delivery solution. You get more bandwidth for your buck at most price points.
- ImageEngine is still the superior option to maximize image optimization and performance. Even better, it requires less effort to implement, and its performance isn’t dependent on other strategies, like responsive syntax. This has the added effect of giving you better performance on a wider range of devices, providing a level of “foolproof-ness” going forward.
— — -
End of part II. You can find part I of this comparison series here.
— — -
