The topic of placing products on the homepage always comes with debate, but when thinking in the context of customer experience design, the reasons against presenting products are straightforward. 

I've previously taken an in-depth look at product page design, and now it's time to turn the spotlight on the homepage.

Purpose of the homepage

The purpose of the homepage is to maintain a consumer's 'buying momentum'.

Buying momentum is characterised by a series of micro-actions consumers take in their journey leading to a macro-action (a purchase).

The team at Site Tuners say the purpose of the homepage is to “get visitors off the homepage”, which is right, the homepage is a “sign post… it’s not where your content really lives.”  

When a consumer lands on a homepage they are at least one step into their journey and want to continue moving closer to their goal/macro-action.  

A consumer may have come from:

  1. A search engine.
  2. A marketplace.
  3. Viewed another form of media stimulating an action (received an email, saw a TV commercial).
  4. A competitor site where they had a poor experience.

Customer’s rarely think, “I hope this retailer has the product I am looking for on the homepage.”   

They expect to take steps, and yet retailers feel compelled to present products.  

Some of this may come from the old school thinking of "three clicks to a product", an old usability rule ensuring products were presented by the third click.  

"Three clicks to a product" still has merit on the proviso the products presented by the third click are relevant and aligned to the consumer's intent.

The ideal homepage formula

For the homepage to maintain momentum and achieve its purpose it requires a cocktail of functionality and content to present four fundamental navigation options:

  1. Large search box.
  2. Clearly displayed main navigation bar/menu.
  3. Mega menu
  4. Content tiles in the body of the homepage pictorially representing main categories. 

Many retailers get #1, #2, and #3 right, however, #4 appears to be the forgotten piece of the homepage puzzle. 

Retailers cannot predict consumer intent when they first land on the homepage 

For those who believe they can present relevant products on the homepage, heed the words of Howard Tullman (CEO of 1871, entrepreneurial hub for digital startups) who was quoted by when discussing the future of retail:

Comprehensive use of demographic data will be useful, but no longer a competitive differentiation.

And even basic "interest" and social information won't be sufficient to win the battle because the new consumer behavior drivers won't be uniform or consistent, even on an individual basis.

Each time a consumer appears on a retailer’s site - it's essentially a brand-new day dictated and determined "in the moment" by a consumer's then-dominant and most pressing desires... intent.  

Google talks extensively about what Howard refers to as being "in the moment”. Google calls it consumer "micro-moments"...

Mobile has fractured the consumer journey into hundreds of real-time, intent-driven micro-moments.

The bigger the retailer, the broader the range of intent driven micro-moments coming your way making the ability to present relevant products on the homepage extremely difficult.

The negative impacts of featuring products on the homepage

Presenting products negatively impacts the consumer experience in three ways:

1. Jeopardises consumer buying momentum.

Because micro-actions have occurred before landing on the homepage, the consumer has established momentum and is trying to get closer to meeting his/her need.  

Featuring products on the homepage has the potential of halting this momentum. Check your homepage bounce rate metrics.

2. Reduces the chance of consumers scrolling down the page.

The dynamic of “The Fold” has changed. Gone are the days of cramming content into the top portion of the screen.  

Content placement is now a matter of establishing a content hierarchy and ensuring the most relevant content is at the top of the page giving consumers confidence there is more relevant content below the fold.

If consumers see irrelevant content at the top of the page (i.e. products), they are less likely to scroll. This is Facebook’s foundation strategy - ensure newsfeed content is highly relevant to initiate engagement.

This from Facebook:

Our goal with News Feed has always been to show people the things they want to see. When people see content that’s relevant to them, they’re more likely to be engaged. 

The approach to the homepage is no different. Scrolling is the new "engagement currency", it's the new pageview.  

Featuring irrelevant content at the top of mobile screens has a far greater negative impact:

Consumers who purchase on mobile scroll 23% more than users who did not, while on tablets consumers scroll 25% more.  

3. First-time visitors will not appreciate the breadth of a retailer’s product mix.  

Baymard Institute conducted a study with first-time visitors and found retailers who featured products on their homepage delivered the perception of having an "overall narrow range".  

First-time users with little to no prior knowledge of the site’s brand will largely base their understanding of a retailer's product range on the homepage content and main navigation categories.  


Toys R Us has products recommended for me. I have never purchased from Toys R Us, nor have I scrolled through the site before (my kids are all grown up).  

Further down the page Toys R Us goes on to say “Shoppers like you also liked”.

How does it know me? And frankly I’m insulted it thinks I would like a vacuum.

Neiman Marcus does little to communicate its broad range of products with this content.


This large retailer thinks I would be interested in patio furniture because it's summer:


Benefits to replacing product content with pictorial content of main categories  

Other than aiding the consumer in moving them off of the homepage, there are four other benefits to this content:

Firstly, displaying main categories within tiles provides a more visual method to move deeper into the site and thus utilises best practice usability and user experience principles.  

Secondly, main category tiles satisfy the brain’s desire to consume visual content.  

Because the human brain is programmed to more efficiently absorb pictures over words, the content in the body of the homepage has greater visual impact on a consumer's perception of product mix offering vs. the main navigation bar.  

In 2014, Marcel Just, Director of the Center for Cognitive Brain Imaging at Carnegie Mellon University had this to say:

Processing print isn’t something the human brain was built for. The printed word is a human artifact.

It’s very convenient and it’s worked very well for us for 5,000 years, but it’s an invention of human beings. By contrast Mother Nature has built into our brain our ability to see the visual world and interpret it.

This content better communicates what main category ranges consist of. For example, the Nieman Marcus “Home” main category title is vague.  

The addition of a tile would better communicate range. Another option would be to include a list of sub categories to further communicate depth.  

Farmers, a large department store retailer, has done this:

The combination of image and sub category links is a good tactic for retailers with larger product ranges.  

Thirdly, this main category tile content will elegantly translate to tablet screens providing large "finger targets" and simplifying the steps to drill down.  

Main category links become less effective targets on tablet screens and the mouse-over effect that triggers the mega menu no longer becomes a usable prompt.

Fourthly, the combination of HTML and images along with the correct category naming conventions assists in the retailer's SEO strategy.

By placing the main category tiles below the feature banner, it helps bring this content above the fold (assuming the banner size is not too deep).  

Farmers homepage: Heat map reporting for this page layout shows significant click activity for all tiles.



Both Walmart and Home Depot have this content on the homepage but it's too far down the page:


Home Depot 


What to do with “New Arrivals” & “Best Sellers” 

For retailers who argue they sell feature products (such as "New Arrivals") when they are on the homepage, consider the customer experience.  

What's better? Cramming feature products into the homepage, or properly merchandising these ranges on dedicated landing pages, then featuring a tile/banner on the homepage communicating the range and its value.  

If the messaging is clear, adding one obvious step still allows you to present products within thre clicks.

The homepage is cleaner, there is more room for main category content, and existing and future customers are given more options. Win-win.

Greg Randall

Published 13 January, 2016 by Greg Randall

Greg Randall is a senior digital consultant with Comma Consulting and a contributor to Econsultancy. You can connect with Greg on LinkedIn or Twitter

26 more posts from this author

You might be interested in

Comments (28)

Save or Cancel
Pete Austin

Pete Austin, Founder and Author at Fresh Relevance

Re: "ecommerce retailers should never place products on the homepage".

Hmm. I admire your bravery. Let me introduce you to a little company, that you may not have heard about, but that seems to be doing just fine, despite putting a shed load of products on its home page:

Here's more about why I disagree with the premise

over 2 years ago


Peter Mowforth, CEO at INDEZ Ltd

I don't agree with this theory. Traditional shops place products or product posters in their windows and have done so for hundreds of (user-testing) years.
The key thing about having a few select products highlighted on the home page is to let visitors know that it's an ecommerce trading site where they can actually buy things.

over 2 years ago

Tim Schwarz

Tim Schwarz, Head of Digital Marketing at University of Surrey

It's always a fantastic topic of debate, especially between eCommerce, UX and editorial functions within a business. More often the case, you are probably correct (especially when you monitor CTR of homepage product promos vs the route to product via the nav).

From experience, the only cases that spring to mind where you can see an improved CTR is during periods of significant ATL promotion of a product that drives high traffic to the company homepage (ensuring the homepage creative has visual connections with the ATL advertising). And also retargeted / personalised banners to users who have previously visited the site - an actually useful quick link to what was previously viewed on the site or recommendations based on their previous viewing habits.

I don' t think products will be disappearing from homepages any time soon!

over 2 years ago


Moises Marquez, CEO at PCX

@Pete Austin, I was thinking the same, but in my case was NewEgg instead of Amazon :P.

over 2 years ago

Greg Randall

Greg Randall, Director at Econsultancy Guest Access TRAININGSmall Business Multi-user


Thanks for taking the time to reading my post and providing your thoughts. It's a very easy argument to merely say "Amazon does it so it must be OK". This is the trap many retailers get into and is why I wrote this post.

A couple points on your Amazon comment:

1. What type of click through does Amazon achieve from its featured products? Does anyone know?

2. Amazon structures its main category pages (essentially mini home pages) with layouts as I have suggested in my article. Have a look:

The content on your landing page does not provide any proof or argument against comments made in this article. What specific point are you trying to make?

Thanks again.

over 2 years ago

Greg Randall

Greg Randall, Director at Econsultancy Guest Access TRAININGSmall Business Multi-user


I like the points you make, in particular, your comments around promotion campaigns resulting in heavy home page traffic. It's important to keep in mind, one of the purposes of the "featured content area" (area sitting at the top of the home page below the header) is to present content that is aligned to marketing mix communications that will stimulate branded searches. All home page examples above contain this content element.

I agree with you, products will be appearing on home pages for a very long time. The purpose of this article is to make sure Retailers understand what they are doing to new and existing customers by presenting product content on the home page. If I help 1 retailer improve their customer experience with this article, this has all been worth it:)

Thanks for you time, appreciate your comments.

over 2 years ago

Jason Buck

Jason Buck, Consultant at The Long Dog Digital

Nice, substantial article with a lot of sources and references.

After banging this drum with a number of clients I went back to one particularly large client's site today, only to see that the rot has begun to set in again with HUGE banners for a single product (also a global navigation called 'Clearance' ... presumably because , unbeknownst to me, their clients have started searching for any old product, so long as it's 'Clearance').
Nice work.

over 2 years ago

Alexander Levashov

Alexander Levashov, Director at Magenable

It's interesting topic and the arguments of Greg makes some sense.

I have couple concerns however with the recommended approach:

1. If we talk about intents, what about people who don't have clear intent (say they need to buy a Xmas gift) and will appreciate some guidance?

2. Having image/picture representation of categories may be quite tricky: say you are multi-brand retailer and sell all kind of dresses as one main category? What kind of dress should you select to represent the category: beach dress, cocktail dress or business dress? Wrong choice can alienate some customers. A word may be a safer option than an image.

over 2 years ago

Pete Fairburn

Pete Fairburn, Managing Director at morphsites

Our experience agrees with your article Greg. Wherever we have replaced featured products with categories as the primary or only homepage navigation, we have seen reductions in homepage bounce rates, higher click through and improved conversions.

Time after time heat mapping and in page analysts backs this up, showing clicks on homepage featured products typically being <1% (often much lower that that) whereas category tile clicks are much higher.

We feel some products based on solid data personalisation may be of interest in some scenarios, but we are still establishing what patterns there are on this.

Happy to share some data insights with you if of interest.

over 2 years ago


Grahame Palmer, Marketing Director at All Things Ecommerce Ltd

Excellent and extremely good advice. Having been in online retail for nearly 16 years, what is being said here is 100% accurate. Our experience in this field bears out all these points - all of which we employ on our sites. And we learned them through experience, so it's nice to see that we've been making improvements over the years that experts like Greg Randall now advocate. Much of it relates to navigational LOGIC and CONSISTENCY. Customers like to see a logical site, where the procedures are clearly evident and which follow a logical pattern. After a few clicks, the structure of the site should become obvious and intuitive to the visitor. There must be a hierarchy in this navigational structure that becomes obvious to the visitor as rapidly as possible.

It's like visiting a shopping mall, where at the entrance, a map of the mall shows where everything is - categories by both product types and retailers' names. So you can NAVIGATE to where you'd like to be rapidly and efficiently.

The drill-down needs to make sense to the visitor, so at each level, they intuitively understand where they are in the site. Category images (and sub-category images) should be different from product images, as this subliminally indicates these "levels" in the webshop.

These all result in the customer knowing how to use the site, and in our experience, adds considerably to the customers' satisfaction. We see an average time-on-site of nearly THREE MINUTES, with average page-views in the realm of 18 to 24. Our bounce rates are consistently UNDER 2% ! On one site, the bounce rate has never gone higher than 1% !

Fantastic advice Greg... I fully agree with all of it

over 2 years ago


Jim Hunter, Consultant at VersionUX

I can see how this might apply sometimes for retailers with large catalogues but even then it's tenuous; with entry points being search engines for example then you'd want to present the product they'd searched on if you had that data . Those with smaller offerings might wish to go with a rule-of-3 type offering on the home page. I think the argument might need to be finer - if randomly 'featured' products are shown and have minimal chance of relevance then I agree this isn't much of a strategy.

over 2 years ago


Grahame Palmer, Marketing Director at All Things Ecommerce Ltd

@Pete Austin
One also has to take into account HOW people got to a website in the first place. Most often, it's through a search engine - and with search engines (read "Google") becoming far more accurate in rendering results akin to what the searcher is looking for, there's now a greater chance that a visitor will be linked directly to product pages, rather than be taken to a "home page".

With large, and extremely well-known portals like Amazon, the principal need for a 3rd-party search engine is not there... people put "" into their browser address bar - because that's the "shopping mall" they want to visit.

Once there - and you'll see it on Amazon - the SEARCH BOX is a headline feature - so in a sense, Amazon is behaving like a search engine to its home page visitors.

Remember too, that "advertising" your product on Amazon's home page provides Amazon with a nice little earner. I would bet that in 99% of cases, most visitors to Amazon's home page ignore this showcasing and go straight to the search box...

Remember too, that with improvements in search engine efficiency over the years, people tend to know what they want BEFORE they actually shop somewhere. By the time they reach your site, they have probably got SEVERAL retail sites across their browser tabs, and are now looking for the site that meets their needs and expectations.

Many don't even get to a site's home page. More than HALF our visitors (taken from Google Analytics data) never touch "index.php" - they click a Google result that takes them straight to the relevant product (or product category)..

over 2 years ago

Joey Moore

Joey Moore, Director, Product Marketing at EpiserverSmall Business Multi-user

Thanks for the interesting article Greg but I’m afraid I don’t agree with your argument after having been involved in many different homepage tests over the years.

I do agree that showing a very limited selection of your range might turn off customers that aren’t interested in those specific products, but there is no reason why that needs to be the case in this day and age.

The flip side of your article seems to argue for a generic, one size fits page that appeals to the lowest common denominator.

I still see too many sites that show the same static page to every one of their hundreds of thousands of weekly visitors, regardless of the customer’s previous interactions or the journey they took to get there.

Even just a basic set up addresses some of most common personalisation opportunities would be beneficial:

• Returning customers
o Items that are cheaper when they last viewed them
o Items that they’ve abandoned in their basket
o Best selling products from their favourite categories
o Items they’ve viewed X times but not purchased
o Their usual shopping gender

• New Customers
o Customer location
o Customer device
o Seasonally appropriate (yes, Patio furniture in summer is fine)
o Did they come from paid or organic search (keyword available?)
o Top selling products from the top selling categories
o Trending over last 24hrs

Intelligent, data led strategies such as these have proven to significantly increase CTR, reduce bounce and earn more £ per impression than generic tiles on numerous tests.

There is of course more that can be done to increase the relevancy for every user. At a fundamental level, we believe that this can be done by analysing and processing all of the micro-actions that a user, and other users, perform using machine learning and other approaches to better understand that user intent.

over 2 years ago

Daniel Grainger

Daniel Grainger, Senior Product Analytics Manager at ASOS

Really interesting article and definite food for thought. Having seen all the various opinions in the comments though, this really does just drive home one major thing for me...there is no such thing as best practice.

"What works for Peter won't work for Paul". It's simply a case of needing methodical and rigorous split testing programme to prove what is right for YOUR company, not everyone's.

over 2 years ago


Grahame Palmer, Marketing Director at All Things Ecommerce Ltd

A lot of what you are saying has relevance... but, much of it is dependant on the system recognising the visitor BEFORE they "login". Cookies are one way to achieve this, but a lot of pertinent data (eg: historical purchase activity) is proprietary, so is only "made known" to the system AFTER a person has logged into an account. Any webmaster configuring the system to reveal this on a cookie is probably acting irresponsibly as it an be insecure! Amazon manages it - but then they have millions to throw at the technology... It's expensive - we've researched it.

So, much of the really valuable data is often hidden from view UNLESS and UNTIL the user logs in.

For new users - or first-time visitors - there is either no historical usage data, or it is scant and largely irrelevant.

Besides, as I have said in previous posts, by the time a visitor gets to your site, they already know what they are after, and have done the preliminary "drilling down" on a 3rd-party search engine, and in an increasing number of cases, the landing page off the search result is product-specific (or one level up, at a category, or sub-category). Our evidence shows a steadily declining number of visitors coming to the "home pages", and we expect this to continue to decline as search engines improve their capacity to show deeper links in search results.

For an increasing number of visitors (first time or seasoned), what's going on on the "home page" is largely irrelevant, as many of them may never see it.

Over the last few years, our focus has been to optimise product pages, and to embed features that the search engines are now hungry for... such as inline meta-data (structured data - the "semantic web", as it's now called). When a product page is well-constructed, both "technically" and in terms of content, it is likely to feature strongly in search results, and will provide the person doing the search with a direct link to the PRODUCT THEY ARE AFTER.

The home page MUST be a "map"... a guide, as it will cater for the (diminishing number of) first-time and/or infrequent visitors who link to the site outside of the common trend (that being through search engines) - and who follow a top-level URL manually provided, such as link someone may insert in a social network posting or blog.

Since optimising PRODUCT pages, we have seen a significant improvement in Search Engine ranking for these specific pages. What's also interesting is to see that visitors who get straight to the desired product tend to THEN navigate around the site, AFTER adding that product to the cart. At least 1 in 5 such visitors buy additional goods. So now we are using cross-sell and up-sell on PRODUCT pages. Already we are seeing an increase in pages-per-visit, and if we get this right (not easy), we are aiming to up-sell / cross-sell at least 40% of visitors, compared with the 20% who do so presently.

Our view is that the "home page" is becoming less significant in the greater scheme of things, as it is increasingly likely that the site's main landing pages (from search engine results) will be the actual products they are looking for.

over 2 years ago

Mark Selwyn

Mark Selwyn, eCommerce and Multi-Channel Retail Consultant

Some great products with great images at great prices giving great value on the homepage can act as a great teaser for the rest of the site i.e. a real signpost, rather than a "we stock jeans" signpost.

over 2 years ago

Matt Naughton

Matt Naughton, Head of Digital Marketing at Lights4fun

hmmmm sorry Greg I have to disagree with this theory.
For us as a niche market, product placement on homepage is vital for our journey. We hold them below the fold (of which less than 25% of visitors see) however they provide valuable education on the range of product on offer (from a niche retailer).
This year using personalisation we're also going to be able to display best suited products to a user too. Utilising dynamic content such as product also makes it much easier for a retailer with a smaller web team to keep the homepage refreshed and interesting for returning visitors.

over 2 years ago

Greg Randall

Greg Randall, Director at Econsultancy Guest Access TRAININGSmall Business Multi-user

Wow, thanks team. I was hoping to start a great debate with this article and have succeeded. Because you all made the effort to read my post and make comments I will make the effort in responding….

@ Jason

Thanks for the kind words and the support!!


Good questions.

Question 1. If we talk about intents, what about people who don't have clear intent (say they need to buy a Xmas gift) and will appreciate some guidance?

Answer. Having category content is the ideal solution for those with undefined intent. In your example, having a “Christmas Gifts for Him/Her” category works. Main category tiles can also be dynamic based on seasonal influencers. Don't think these tiles are carved in stone.

Question 2. Having image/picture representation of categories may be quite tricky: say you are multi-brand retailer and sell all kind of dresses as one main category? What kind of dress should you select to represent the category: beach dress, cocktail dress or business dress? Wrong choice can alienate some customers. A word may be a safer option than an image.

Answer. One option is to have more than 1 model wearing more than 1 dress style. I would also display sub category links (see example above with the “Home” tile) to communicate to consumers the depth of this main category. As I mentioned above, images within the main category tiles should be constantly changing to reflect the varying seasons or buying influences which bring consumers to the home page.

Another point worth making is, main category tiles can vary in size and placement based on their impact to the retailer. If a retailer has dresses making up a large proportion of their sales, let the tile placement and the space it consumes on the home page reflect the same. Another option is to split out the main category and break out dresses.


Thank you for sharing some of your experiences. I would be very interested to hear more about your tests and the results. I have been applying the above methodology for years over hundreds of retailers, and the same thing happens each time:

1. Significant reduction in bounce rates
2. “Page value” for home page increases
3. Increased pageviews per visitors for those land on the home page
4. Increased engagement for those visitors who initially land on another page and go back to the home page

Point #4 is a journey/experience retailers do not account for. When consumers land on a deeper landing page (not the homepage) and find the content as being sub standard or the visitor becomes lost or their buying momentum stalls, what do they do? If they don’t leave the site, they head to the homepage to start again. This main category content helps consumers “re start”. If a retailer thinks they can present relevant content on the home page on this type of experience, they will get it wrong.


Thanks for the support and the brilliant commentary on your experiences. I too take this navigation hierarchy approach for all the reasons you mention, plus they become great landing pages for acquisition strategy.


If consumers are searching for products they should never be landing on the home page. As mentioned above, the home page is receiving generic/branded traffic where there is no clear sign of intent. A consumer may have intent, but if he/she has landed on the home page they have not yet identified it. This is common for retailers with strong brand equity. This consumer type assumes the retailers can accommodate their need so he/she first gets to the site, then begins defining intent by their micro actions.

@Grahame (again☺)

I back your statements RE the search box and Amazon. In my analysis of my clients who are various sizes, there is a correlation between size of product offering (product breadth) and the use of the site search box initiated from the home page.


I am all good with “agreeing to disagree”☺ I would like to address some of your comments:

In your assessment of what type of content you can display to returning visitors (not necessarily “customers”) you are making an assumption their buying intent is based on history, or what they had in a cart when they abandoned. This approach may have been effective 5 plus years ago, but as I mention above, this is extremely hard to get right today when consumers are driven by “micro moments”.

I would like to know more about your approach to presenting relevant products to new customers based on their location and device.

If consumers come from paid or organic traffic sources and have specific keywords identifying intent, they should not be taken to the home page.

You mention using “machine learning” to understand intent, what do you do when a consumer has not defined or identified their intent in the context of “micro moments” (not micro actions)? I agree micro actions will deliver context, but “micro moments” is completely different.

@Grahame (again ☺)

This is great stuff. The point you make about the home page becoming less significant is an interesting one. I only pause in supporting you on this because I think about retailers with strong brand equity.

Coming back to a point I made earlier, the home page does and will continue add value to those consumers who head back to the home page (for whatever reason) and wish to start again. Main category content will help.


Again, happy to agree to disagree. I couldn’t help myself and had a look at the Lights4Fun website. I challenge you to conduct a test. Keep your existing home page and get some heat mapping software on this page to measure click activity before the test begins.

Create a “test homepage” with the following page elements:

Feature image (that can stay the same but ideally make it a bit more shallow)

Create main category tiles and place them on the test homepage with the following visual characteristics:

1.Make them larger and three across the page with bigger images.
2.Contain the tiles to make it easier on the consumer’s eye.
3.Include all primary main categories (indoor lights, outdoor lights, connectable lights, solar lights, wedding lights, dome lights,) – exclude Christmas and accessories for the purpose of this test.
4.Within the tiles, list approximately 3 to 5 of the most popular sub categories. Make them hyperlinks linking off to the sub categories.
5.Have a call to action to view the entire main category range.
6.Keep your best sellers on the home page but push the best sellers below this content.

Give the test the proper amount of time (be weary of sample pollution) and let me know how you go☺

over 2 years ago


Catch Any, Owner at catchany


thanks so much for this post/information/opinion - it is so relevant for my current project. I also appreciate everyone's comments as they to are very important to the discussion. This is a VERY IMPORTANT subject so I hope others will continue to give their feedback.

over 2 years ago


Grahame Palmer, Marketing Director at All Things Ecommerce Ltd


Looking at your challenge to @Matt...

We did just this... several years ago! Largely out of desperation than by design - trying to discover the best ways to increase average order values and get customers to choose products that were more highly priced and with fatter margins, thereby getting a better ROI.

After giving the "test" home page a structure similar to what you describe, we gathered up 20 - 30 family and friends and asked them to visit the OLD site (hidden on a private IP), then to visit the NEW site - and tell us what they thought. All found the new site sensible and logical - including my 84-year-old father in law who was quite "new" to computers at that time.

After unleashing the new site onto the pubic, there was a significant increase in pages-per-visit. This more than doubled, literally overnight. As this was done prior to many of Google's clever "funneling" analysis tools, we weren't able to easily see the chronology of page visits, but we did discover a bit of back-end software that kept track of Sessions, so we installed this and were then able to watch individual customer's movement around the site. This told us the landing page, date and time of arrival, IP address, browser type (user agent), and then each click-thru, along with how much time was spent on each page.

We looked at several hundred sessions over the next 3 months, and we discovered a lot of common patterns (too many to bore you with in this forum) which then provided pointers for us to improve the visitor experience.

And over the course of many months, the tweaks and changes we made (again in a REACTIVE sense, rather than a pre-determined plan) resulted in a structure you advocate.

So we did not start out with a plan for optimal functionality - we used the data to help us improve things... And today it matches your strategy almost perfectly.

Of course, along the way, an incredible amount of change happened - and continues to happen, but I think the thing we learned mostly is that it's the BEHAVIOUR of the online consumer that needs watching most of all. And as technology advances (mobile and smart-phones) so this behaviour adapts to it in ways that suit PEOPLE, not the machines being used.

In addition to doing a lot of work with Google's analytics tools (and the various testing tools), we now make alterations and improvements almost entirely on what the patterns of behaviour are telling us. In many cases, this is driven by technology (the need now to have device-responsive programming and page templates), and in other cases is driven by site design structures.

On top of all of this has to be good marketing "nous". Once we've got a customer's attention, how do we get a good "share of his wallet"?

... but that's a topic for another discussion!

over 2 years ago

Steve James

Steve James, Head of eCommerce at giffgaff

ECommerce retailers are after the incremental, so by selecting a single family of products, like cameras, it's straight forward to demonstrate the incremental WoW uplift for said cameras with a flash sale. However if sales of TV's have reduced, then questions should be asked. "At what cost....?"

The true balance is an uplift of all categories, win-wins, and responding to customer needs. CTR is a good metric for home page.

But it's not always that simple. Perhaps the eCommerce retailer has a surplus of stock of cameras that's burning a hole in warehouse/stock holding - so a flash sale from home page whilst tolerating a slight dip elsewhere, can be a good response from an eCommerce team.

over 2 years ago


Darren DeMatas, Writer at Selfstartr

Wow! Epic article, you even provided some free consulting to a commenter. Bravo!

I tend to think this is right, except for niche ecommerce sites. If you only have 5 products, yes they should go on the home page. You really have to test it though.

over 2 years ago

Laura Hogan

Laura Hogan, Head of Search at Ricemedia

Some definite food for thought here, thanks Greg.

I can see the rationale for both arguments - maybe it's more based on the size of your business and offerings.... If your business has one focus area then having products on the homepage may be of value, but when you're a retailer with many different categories like your ToysRUs & Amazon then showing products can turn a user off (especially when they assume what you'd like.... I wouldn't like a toy vacuum either!).

It'd be good to see an a/b test of the same homepage with and without products to see how clicks differ..

over 2 years ago


Zeljko Simic, Senior Designer and Full-stack Developer at Freelancer

In the same time you are right and wrong. Mostly wrong. Senior graphic designer, web designer, full-stack developer and lecturer here, with experience working for private companies, retailers and also education institutions and public organisations, and also administration officer.

Your theory is easily crashed in reality. And this is how. I did a lot of AB testing (in your post you did not mentioned not even one audit or test) and monitoring and audits over the years and this is what I know from my job experience.

1. Every web shop needs to be designed and developed according to the industry, mission and goals of the company and branding in general. That means that small retailers with few products can be very creative and allow them-self not featuring their "add to cart" products on homepage, and they will get a decent boost in sales, mostly because they are specialized their business for specific type of customers, they know who their buyers are and they are sticking to that group of customers and ignoring the rest. But when you have retailers that sell a lot of categories for people that range from 1 y.o. to 100 y.o. for all sexes, then you need to feature out your products on the homepage, with adding bestselling products and what is on sales at the moment. (I agree with not adding products on "above the fold" section, but after the fold - I agree that above the fold should go image with category AND TEXT + LINK.)

Why? Because bigger retailer will be losing orders and customers, because people when going to general store already know what they want to buy but you also want to offer them something more that they could see and maybe buy it, which happens a lot in general stores of big retailers, but not in a stores of small retailers. You go to buy something in general store and you end up buying 3-4 more things. And forcing people to click on those kind of web shops 3 or 4 times to get to the product is insane, you are just raising the frustration and making your visitors edgy. Why? Because if you are visiting web shop from a mobile phone, average load time of one page would be around 10 - 30 seconds. Now imagine when you need to click minimum 3 times as you say is a better option. That is a lot of waiting, and imagine when visitor wants to scan for other products, that's a hell of a waiting. So your theory does not have practical use. It would actually put big retailers with wide range of products in a big problems.

2. You can't focus yourself only on UX psychology , and ignore marketing, economy rules, psychology of shopping, colors, money income per buyers and so on. So again, your theory is not good, and 8 years ago I was thinking the same as you, but things do not work like that, most people do not understand basic stuff on how internet and marketing works, and your ideas will not work for them - and that means bad conversion - also bouncing rate is relative, it depends on a lot of facts, and serving it like it is something so basic, it is not, because bouncing in most cases depends on websites architecture, and I don't know if you had ever done AB testing, there is a lot of cases where bouncing was high but conversion was also high - as I said bouncing depends on a lot of things, it is not counting only those who left - do you know how much there is bad implementation of Ajax and other things?! Bouncing is not just a leaving ratio.

You need to get into the mind of an average buyer and see what they need and expect, what are the missions and goals of the company, how they want to position their brand on the market, who are their customers, are the customers random people or there is a specific group, if there is a specific group that company is aiming for then be my guest, let your creativity flows, but if they are random people, stick to the standards. And trust me, for big retailers 3 clicks or more and 10 or more big images on homepage is a suicide. Small retailers as I said can be creative about their homepage, and it will work like a charm, but big ones, will lose money and buyers online. Run audit between website performance, UX and average people psychology, it will always be the same. Your theory does not hold.

PS. You were comparing big retailer general store web shops Harvey Norman with small retailers that are specialized them-self for specific customers that by specific product lines. Those are two different approaches, and I wouldn't advise anyone to follow anything you said except the part about placing category images above the fold instead of products. Everything else is just wrong, starting with examples you used - you are confusing apples for oranges.


2 months ago

Greg Randall

Greg Randall, Director at Econsultancy Guest Access TRAININGSmall Business Multi-user

@ Zeljko,

I have a few responses to make….

The data and research methodology:

I am not sure how many organisations you worked with to gather your comments, but this research has come from the analysis of..

1. The hands-on deep exposure to over 500 retailers of various sizes and across all industry types around the world

2. Analysis of hundreds of millions of sessions

3. Data sets coming from a mix of AB testing, heat mapping, and consumer path analysis, and page value metric analysis

Utilising and basing your statements on AB testing in isolation, is a dangerous game to play when you are trying to determine the value the homepage contributes in holistic consumer journeys, not just the next step from the homepage.

Retailers need to know what content (and functionality) mix adds value to the consumer's JOURNEY.

One example of this was found in the consumer path analysis research where consumers come back to the homepage in the middle of his/her journey to gain a frame of reference and/or start over (for various reasons). The purpose of the homepage is NOT just to deliver a “first impression”.

It's multifunctional.

Testing the homepage in an AB environment that is in the wrong context will provide the wrong results. This is why consumer path and “page value” analysis plays a crucial part (page value analysis is the process of attributing a “monetary value” to the homepage as a benchmark, then measuring how this benchmark is influenced when changes are made).

Each time a retailer has made changes to the homepage aligning to the recommendations in this research, the page value increased between 80% to 175%. This is across multiple retail verticals and retailers with various sizes of product ranges.

These results worked for multiple retailers who sit under 50 products on offer in total.

This is not theory. Your statements around me being “wrong” are inaccurate.

If you have data to verify your approach to be effective, please post them. It would be good to see.

Pageload speed:

If you are a retailer who has pageload speeds of between 10 to 30 seconds you have more issues than what’s being presented on the homepage.

No consumer will go beyond the homepage with pageload speeds of this nature, making your point redundant.

In fact, the slowest homepage pageload speeds recorded, are those with products due to their need to call the database.


I would like to address the following statement you make….

“And forcing people to click on those kind of web shops 3 or 4 times to get to the product is insane, you are just raising the frustration and making your visitors edgy”

This statement suggests you do not understand the foundations of user experience design. I suggest you research “interaction cost management” and the importance of balancing the reduction in both mental and physical effort.

I hope this further clarifies things. If not, I have a book coming out later this month solely dedicated to the best practice homepage. I can send you copy.


Greg Randall

2 months ago


Zeljko Simic, Senior Designer and Full-stack Developer at Freelancer

1. If you did research over 500 retailers over various sizes and you did not create a groups of your research, and you generalize like you are doing now and in your article, then your research readings are not usable because you do not understand the difference between apples and oranges. Why? Because you don't have a black-on-white data that exactly says that for this kind of retailer in this and that kind of industry this is what happens. This is common mistake in reading a research data among marketing managers in Ireland, Croatia, Slovenia and so on. I doubt that it is much different in other countries. And in most cases in this countries most marketeers give average or bad advice when building or developing e-commerce site for their clients, which often leads to average or bad sales results, with even worse web performance and optimizations. because with your advice, every retailer would need to invest every year into re-branding or redesign to keep the visitors coming - that is veeeeery expensive. But what is the connection with sale and bounces?!!! Let me show you how wrong you are with your generalizations. A question:

You have two e-commerce sites - one with categories above the fold, and products after the fold mixed with few banner categories. The other one have only categories on the front page as stated in your article. The first one have lets say 45% bounce rate, the other one with categories have 28% bounce rate. Page views for the first store is 4 pages per visitor. For the other store, it is 7 pages per visitor. For first store, average visitor spends 5 minutes on site. On the other store, average visitor spends 9 minutes on the site. First store has 1000 visitors per day, and 60% visitors buys their products, 40% are the returning visitors, or leavers, and everyone else. Second store has 10 000 visitors per day, but only 10% makes a purchase. So this is one of the most common problems for online stores. And you generalize and base your advice only on bouncing ratio. This is data from compering two general store retailers. They are from the same category of retailers, not like you did, by mixing the small fashion retailers with big general store retailers and so on. You know why first store have higher conversion rate?

A tip:
Because they have samples of their products on the front page, so visitors brain after scanning the whole front page content knows what to expect when he goes deeper into the web store, and he stops thinking about the design and focus on products, because in his mind the visual appearance of the product is carved into his memory at a very start and learning time is short, while for the other store, visitor must take time to get to know what to expect, because every page is different and brings more visually different content that can sometimes feel like it is hidden, and that is just a few reasons why second store will need to invest every year into re-designing if it wants to stick to your advice in your article by having only categories - will their sales grow?! Probably not much as expected, but their bounsing ratio will be cool. While newest research from 2018 says that mix of both keeps the front page to feel more alive and friendly which brings more conversions - Amazon is mentioned as a store that uses good principles (but it has average visual organisation) - but that goes against your advices in your article?!!! So, I guess that a research from 2018 beats the research from 2016. Because this what I am writing is what was known before 2016, and now we end up that research from 2018. confirms that research from 2016 is no good. Research from 2016 wanted to bring some new trends, but that is the problem, the trend is never a permanent solution, because trend means generalization, something that everybody should copy - but then it passes. Like it happens with eBay at the moment - eBay will probably be forced to AGAIN redesign their front page and I feel they got caught in some bad loop of bad professional advices.

For example. Lets take eBay. eBay applied literal understanding of the research you talk about, and I will say once more that after your reply, I have big doubts that you did read the whole research, because of constant generalization in your writings, or you do not understand what you did read in the research, and you are generalizing again. But lets get back to eBay, who did literal approach to what you are saying. First few months they got a boost of visits, which happens every time when one retailer does re-branding, or redesign of their e-commerce website - but how bounces are related to sales?! Well, they are not 100% correct, and they take very little importance in GENERAL data. Taking data from those first months will work only for those few months. So we can't generalize around that. This is first thing that makes your article not valid, and it is common mistake for most of digital marketeers to get it wrong because you focus yourself only on newest data without even looking at financial aspects for past few weeks, months, years. Because after those few months, comes the real data. And then you can see if the website you built is usable and practical by comparing it to previous data sales over the weeks, months, years. So, eBay, that applied in literal way what you are saying is experiencing now increase in bouncing rate, decrease in page views and time on site, and last few months their ratings are going lower and lower and when we look at the past few years their sales did not increase so much, year by year they almost have the same percentage of income - so it looks like their new design did not help them much, hm but you say it should (like for more then 100%?!!!), but eBays data says it didn't - so?! While Alibaba which uses Amazons approach but with far better and cleaner organisation and representation of their products in combination with above the fold banner categories has increased their ratings, and their sales actually went for 50% up, which is 43% more then eBay - hm?!!! But you said that that should not boost their sales?!!! Yeah, but you were generalizing. Is it possible that you are wrong?! Yes, big time. So first few months after re-designing their (eBay) website to include only categories, and kick out all product listing, their ratings are dropping. Why? I can mention at least two reasons - 1. number of clicks to get to what you need (on mobile devices that takes time even more then on desktop computers) - considering that eBays e-commerce website for desktop computers is not optimized at all, but desktop computers use high speed internet connection and connection is mostly stable, while mobile data connection even when connected through wi-fi is always very unstable connections which can lead to more waiting. - 2. Lack of visual appearance. Their landing page feels empty, not alive, it feels like a catalog that is missing some energy, and visitor don't know what to expect so their brain can't focus on products, their brain focuses everything, and that means that visitor is stressed. There are more issues, but this is the first that comes to my mind after looking the data and their at the moment look of their stores.

2. If it's multi-functional, then why it doesn't apply to Amazon, Alibaba, AliExpress,.. and so on and on and on. So it is not multi-functional. And here is one of the reasons how having only image categories without few product listing on front page can be frustrated because after a while people need to read, they have urge for textual content. I will be using one of the statement of visitors during one test. So lets imagine a specialized store for tools, house materials and so on. The visitors comes to online store and wants to buy an electric adapter, and at the moment they don't know how it looks - and the categories are tools, batteries, electric charges, and 3 more electric something plus 10 other categories non-related to what visitor need (over 50 products). You know what you just did with that kind of approach for that online store, you just gave a bunch of options and overflowed your visitors with images without any text, and with every click on the category they realize they have more and more problems finding the product because there is just too many options from the very start and not much of the products. And they leave after 20 minutes. So you got low bounce, you constructed a website by forcing people to open more pages to boost your web ratings, which is calculated into your Page view stats, and you got 10 minutes per visitor on stats for Time spent on site - well of course, and yet the visitor leaves - you get nice stats but zero money. And with all that, you failed to convert the visit into the sale, because after all that time, your potential buyer just left the store, a little bit more annoyed and frustrated. So if that is multi-functional for you then ok, in real world this is what happens. So at the end it can't be multi-functional if that happens, right?! Yes, you could gather all electric categories and so on and make them as sub-categories by creating one big parent category - Electricity - but then ask yourself about your multi-functional theory, would you be willing to click on banner of parent category, then on banner child sub-category and then on listings, then on filter, and then on product to actually read about it, and lets say you need to read and compare few of them - WOOOOW. Average visitor already lost himself inside your tree of categories and subcategories and it just becomes too hard to visualize them-self's where are they now. And you know what, people just leave when they end up in that phase. You get your high stats, with 0-25% sales.

Example: Amazon sells to 10 countries, while eBay in over 20, eBay followed your advice of including only categories without products, yet their stats are dropping fast. eBay has lower bouncing rate, Amazon have higher. Yet Amazon have higher conversion then eBay for aprox. 60-80%. Multi-functional?! No, it is not.

3. I learned this from my years of work and studying:
1. General data does not apply generally, it is used as a starting point that helps researchers to compare their own data and see in which case and when NEW research data can be applied. So, I am not applying that you are wrong, I am saying that you are wrong.
2. Trends are not multi-functional, nor the general researches.
3. AB testing consists of (heat mapping, and consumer path analysis, and page value metric analysis and much more) - they are not separate things, or stand alone, all for them-self. Professional Digital Consultant should know that.

4. About stores load time:
Do tests on mobile data for all major retailers (please do for your clients also that you have on your website - everybody will see that practices about load time you speak off here does not apply to them - also read google developers blog about the same issue, there is a lot of statistics there) for every type of the connection, and see how much seconds for every online store it takes. Wait, I can run audit of for example Farmer that you mentioned. Here comes less then basic test done in 3 minutes:
1. Desktop - wifi - 15 seconds
2. Smartphone - wifi - 13 seconds
And this is withoud even testing for 4g, 3g, 2g and edge. And you said, quoting:
"If you are a retailer who has page load speeds of between 10 to 30 seconds you have more issues than what’s being presented on the homepage. No consumer will go beyond the homepage with page load speeds of this nature, making your point redundant."

But they will. This is the problem, you are only a Digital Consultant, you have no knowledge of design and human psychology, you do not understand the organisation of the content, the story, you do not know how to use the text, you do not know how to predict some actions, you do not understand the behavior of people that are using the technology, and that is why you don't have nothing to compare your digital data with, and that is why your view is so small.

Hmmm... so this shows again, how you are saying one thing, but data comes in very different way, also let them know they have 6 big code errors and they use a lot of depricated classes and functions which slows their loading time drastically. And you say with your help they improved their business and conversion, but then you said people would leave such sites. This shows me how much you don't understand.

5. You say, quoting:
"Each time a retailer has made changes to the homepage aligning to the recommendations in this research, the page value increased between 80% to 175%. This is across multiple retail verticals and retailers with various sizes of product ranges."

So, if it increased, and if it is generally multi-functional for everyone as you stated in your comments and article, how does eBays change did gave them only less then 2% increase in sales? This is public data, feel free to get their financial stats over the past few years before they applied your theory. How does companies that still uses mix of banners + products listing on their homepage still have higher conversion. I mean, yes, you followed clicks and everything else which I explained in details how those two behave in real life, but how can someone with landing page as Amazon have 43% more increase in profit, then company that did apply your theory?!!! Hmmm.

6. Quoting you:
"These results worked for multiple retailers who sit under 50 products on offer in total."

But you said also this for your research:
"This is across multiple retail verticals and retailers with various sizes of product ranges."
So, you love to generalize, which often leads to false data presentation, and you end up negating what you state, which is very contradictory, and being inconsistent raises more doubts in quality of your work. Do you agree? I do.

I remember you said your research is multi-functional?! So it is or it is not multi-functional anymore?! I mean you were comparing small fashion stores with big general stores with ease generalizing everything, but now you say that your research only worked for multiple retailers who have UNDER 50 products. And I also said, that small retailers can go creative because it can help them to build an image on the market. In your article you generalized, and some of the experts here pointed that out and I agree with them and at the end we see that you are very very very wrong, and you still continued to point out how your research is multi-functional, and in the end of your own comment you said that only small retailers witnessed the success, but I am pretty sure that only few of them did, while others are still fighting to maximize their sales. Considering how well you are at generalizing the issues and problems. I alone cover 5 phases of development, design and management (I don't depend on managers, designers and developers so you can't bull-shit to me), from first client talk to the implementation of the solution. I worked on all of those positions, I know the process, and when someone who specialized only one small part of development process (digital consulting) says that something is multi-functional, I will bring the whole army of examples where it is not.

All data I mentioned can be verified, after all it is public.

PS. Your companies website loads very slow, images are too big, load time on high-speed internet is over 1 minute, - over 2 minutes on smartphones, and you used the worst possible practice in writing your about page - which makes your website extremely expensive to visit. Talking in third person removes all colors from your character and you get more people leaving your site then staying and continue reading no matter how good your images are - just pointing out psychological elements from designers perspective that harms your website. Your website is too long and there is too much of you in every section which sends the message of ego-centrism and removes friendliness. So...

If you can't solve your own website, how will you help others?!!! Your website should be an example of your practices, and right now it is not even close. Don't take any general data literally, because those data is not multi-functional or ment to be used literally.


2 months ago


Nicola Hasted, Freelance Content Writer at Freelance Writer

I'm looking at ways to help an e-commerce client of mine and this is a really useful article. My client is only a small business and I don't think they have the breadth of products to warrant the category format. However, I think their current layout focuses too much on one product at a time (large slow, slider) so still needs to truly represent all that they offer. Anyway, I'm new to this but one thing I do get is that Greg knows how to format text for online consumption and Zeljko - not so much.

20 days ago

Greg Randall

Greg Randall, Director at Econsultancy Guest Access TRAININGSmall Business Multi-user


Thanks for the great feedback. 99% of the time I receive great feedback from people like yourself. Then there are the 1% who are those suffering from the Dunning-Kruger effect:)

All the best.


20 days ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.