4 Major Challenges Behind an e-Commerce Review Aggregation Tool

Tuesday, November 20th, 2018

 

One would say that bringing all product reviews together, in one place, is just as simple as doing that work manually, the only difference consisting of dramatically reduced time on the operation due to automation. That means that a robotic process replaces the exact tasks that a human being would otherwise perform in an Excel file, copying and pasting all review details, one by one  – approach that is still being used today by from small to large brands, too, especially outside of the US market.

Less time on this job is definitely one of the biggest advantages which becomes more and more important when the reviewed product is far from being a simple one. For example, in case of products that have more variants – a wide range of assortments, rich brands covering numerous product lines or when sales and distribution take place through many e-tailers, the review aggregation becomes quite complex.

 

Automation makes review aggregation seem fast and easy and at 1-click distance!

 

While this is the perception that everyone outside this business gets, the behind the scene work is by far more challenging.

Here are the top 4 challenges that we have successfully overcome:

 

1. Review Screening

With a growing eCommerce world both in terms of number of online transactions but also online retailers, every eCommerce website wants to offer their visitors the best possible user experience to support sales conversion and returning customers. That comes with on-going tests which often impact how these websites’ content, including reviews, is displayed to consumers and how their APIs deliver the review data, wherever they even exist.

Like any consumer who visits an eCommerce website and needs to figure it out how to search for a product and read its reviews,  sometimes getting confused due to website recurring small changes, a robotic process also requires the right automation which needs to continuously be maintained to follow websites and APIs’ (late) changes. Some of the frequent changes are caused by:

  • how reviews and ratings are displayed on a product page
  • whether all review pages (as reviews are organized across multiple pages) are accessible and where they’re not, we need to identify the cause and the screening solution for that particular case
  • on-page changes

And if you are asking about APIs then you have to know that APIs are either not always available or not always working the right way. This is a consequence of the fact that consumer experience is prioritized against consumer goods companies’ whose products are sold on those websites.

Therefore, we had to define the best possible review data structure which is essential in order to  create a reliable screening mechanism that simultaneously meets three conditions:

  • it is valuable for all our customers from large to small consumer goods companies
  • it meets (ideally) all websites’ structures
  • it is extensible enough to adjust to any new eCommerce website and changes of the current websites

In order to ensure an accurate review scanning and aggregation service, the resources behind could be extremely large – if we only think of Amazon, a marketplace that we are supporting, which has billions of pages.

 

2. Review Aggregation Speed 

New review discovery needs to be done fast and that is possible by engaging different infrastructure solutions depending on our customers’ timing needs. The usual required timing is once per day and with more sophisticated infrastructure solutions the timing can be shorten to twice per day. Yes, we could optimize that even further, however one needs to assess whether highly optimized aggregation timings are really necessary and what are the actual outcomes when having them in place.

 

3. Perfect Product Assortment

Following the well-known principle “garbage in, garbage out”, in order to get your product reviews in good order you need to ensure product data consistency across all sources: eCommerce sites, the review monitoring tool and potentially your internal systems as the first source of evidence of your products and their assortments. For that, you need to make sure you define a common terminology for what a product means across all these places.

If we look at our customers that are typically companies that design, manufacture and sell goods such as electronic devices, apparel, footwear, baby care products and many more, we see that all their products are available in a wide range of assortments. For example, if we take footwear and consider color X size, then the assortment variety is quite significant and can take us to tens of variants for a single product. For all that we get the reviews right! Therefore, with FeedCheck we support review aggregation and analysis at 3 levels:

  • Assortment level – which corresponds to every product link 
  • Product level – which corresponds to all assortments available for a product
  • Group level – which may correspond to a brand, product category and any other pillar relevant to your business

In order to get the reviews right, we help our customers with the entire product and product assortments mapping and standardization so that they’re able to know exactly what products they receive review alerts for and how to relate the review and rating numbers.

 

4. Unique Review Content (or No Duplicates Allowed)

Aggregating review data while excluding duplicate records was another challenge on our journey which came accompanied by a series of questions:

       How do you know if a review is new or just an old one edited by its reviewer?

       How do you know a review will not get duplicated due to …unknown factors?

      How do you know how to distinguish between 3 reviews, all of 5 stars, having the exact same title and content?

      How do you correctly link each review to its actual product variant, in case a product is available in multiple assortments?

…and many more which invaded our brains like an avalanche as we were developing the FeedCheck platform.

Having navigated all these and finding optimum solutions, FeedCheck ensures today unique review records which is a must-have requirement for any kind of analytics one would need to perform on their reviews.

 

We hope that by sharing from our major challenges you will learn how review aggregation is possible for you too and get the comfort and trust that a platform like FeedCheck is walking on a beaten path to bring your reviews from everywhere in perfect order.