We are developing a saas program for the ecommerce platform. Each product has a name, price, sku, description, variant and etc. By going into a product page, users can edit each product. Instead, our developers suggest we just have a huge table like excel and they can input the data. Their reasoning is it works and thats how people can manage a huge amounts of data. This is also suggested for the variants as well.

For variants, I had it where users will need to predefine their variants before assigning them to products. Eg, variants of color will be red,blue,green, orange. Size will be 8,9,10,11,12.

What the developers are saying is that predefining the variants is unnecessary cause u can just punch it in the table.

Its true, the table works. I don't know how to justify my reasoning for why its better to predefine variants and having a product page. Can someone explain to me why it still feels disjointed and dull even though the new table design works?

  • There isn't enough information to give an opinion. However you needn't emulate a spreadsheet. There are numerous ways to solve this problem,
    – Mayo
    Sep 16, 2015 at 20:47
  • 2
    It's true that a lot of UIs would likely have been better off implemented as spreadsheets.
    – DA01
    Sep 16, 2015 at 21:01

3 Answers 3


Seems like a question of user workflow.

Question to ask: Is it more likely for users to go in and just add/edit 1 or 2 products? Or is it more likely for users to go in and make a lot of changes?

A spreadsheet format works best for heavy data entry and the user have on hand all the info they need to make edits.

Editing on individual product pages make sense, when the user only have to make a handful of updates and need to decide what to edit based on the info on a particular product. E.g. Should I price the product higher or lower based on viewing the demand/order history data on that particular product.


The cons of tables

In addition to what nightning said, consider:

  • With tables all the data is in one raw, which can be an issue if there are many columns (as your developers may testify if they ever used database admin apps).
  • Tables can't support hierarchy, which means the design won't scale.

Best of both worlds?

So you'd be losing some visual clarity, structure, scalability by choosing tables. Sure, if users made many frequent changes a table can be a nice solution.

You may also wish to consider a hybrid - editable table and a form. Or collapsed rows that reveal a form.


I think both you and your developers are right at least to some extent. It's undeniable that using tables to modify lots of items is a fast approach, and whoever used Excel or the likes will use this easily with very little friction. So they're correct in terms of usability flow.

However, there's an aspect you seem to be considering and they don't: information architecture. Pre-defining subsets of data not only is a good thing, but it is absolutely necessary, since letting your users freely add content without any guidance or limit is asking for trouble. A simple and stupid example to illustrate this point: let's say your product's variant taxonomy is color. You know your product has 3 possible values: blue, red, pink. Then, instead of define variants, you let your users do whatever they want, and you start to find Unexpected vLues like cyan, magenta, fuchsia or whatever. Now, you know you have 3 values, yet you actually have 5, 6, 10 or whatever. This is because you failed to provide a valid taxonomy that works as a guide.

Also once you work on IA, which should omes first and before anything, this will dictate some logical approaches and procedures. For example, now you can use tables with cells containing dropdowns with pre-defined values, which is way better and way faster than having your users write the values (let alone the benefits at the time of data analysis). Also, it's possible that after a proper and thorough IA study you realize your best approach is to use cards, or whatever your analysis dictates.

In short, the problem here comes because of lack of definitions and structure. Try narrowing your data and you'll see patterrns emerging, which in turn will tell you the better approach to use. An easy way to do this: use a table just like tour developers say, and once you see recurring values, you'll recognize patterns and taxonomies to improve your app's usability

  • "...letting your users freely add content without any guidance or limit is asking for trouble". This sounds right in terms of how i feel. But aside from users not knowing what values they input, are there any other troubles that will appear?
    – Bearsaur
    Sep 17, 2015 at 20:01
  • of course. You have 2 sides: the end user (for example visitors or employees) and the admin user(s). You can see how the end user will have issues due to inconsistent data management, now imagine the burden it will cause to the admin user(s) or those trying to analyze that data. Also, think in terms of usability, what is faster and better? To write a random word or to have the word you need to choose already displayed in a list? So basically we're talking of speed, data rationalization and ease of use
    – Devin
    Sep 17, 2015 at 20:08
  • Would a separate page to manage variants be beneficial or what dev suggested, when values are typed into the table, they automatically save and admins can "recall it" with auto complete (like how google auto completes your sentence). So extra step for user or freely add content that might result in chance of error?
    – Bearsaur
    Sep 17, 2015 at 20:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.