marc_firedrop

I can't share the set itself but can tell you a portion of it was made up of scans of open-source website template that we've hand picked based on popularity and quality and then labelled using Crowd Flower. The minimum we are looking to achieve is websites that can compete with anything you might find on Themeforest or Wordpress.

Thanks for the kudos! Hopefully we can deliver where The Grid couldn't. We actually paid close attention to what they did and think we've learned a few things.

The main difference is that we are spending more time learning what the user likes first before offering design options. Options being a key word here, because The Grid doesn't really have much of a feedback system and even their latest V2 alpha resorts largely to guesswork, whereas our system will give you choices and ask questions like "Do you like this?" at the appropriate points.

The experience we're going for is the one you get from working with a human web designer: at first it's a two-way discovery process and then, over time, the web designer gets to know and understand you and therefore becomes better at nailing what you want straight away. Simultaneously, the platform will learn from anonymised behaviours from its other users, mixed with analytics monitoring, to discover recommendations that it can pass on to you as a user. For example, imagine you're a restaurant owner and you have a button that links to your booking engine which is labelled "Reserve a table". Our platform learns, from other restaurant sites on the network, that "Book a table" has a higher conversion rate, so it can pass that recommendation automatically on to your site. That kind of thing.

That's actually spot on. I ran digital agencies for the past few years and it was always important that the client had some kind of input to the process, regardless of how good the design concept was. It's a simple pyschological need for a person to feel some kind of ownership and connection with their site, which is where I really feel The Grid has gone wrong because their designs are so far off what people want or are used to.

We're updating the site in the coming days to demonstrate better how we involve the user, and presenting different layouts is actually one of the ways in which we do it. Firedrop is more of an A.I. web designer than a tool: you interact with it in a much more engaging and human way.

I think the idea behind The Grid was good but perhaps the execution was not so great. We're aiming for a better execution of the concept, with more of a focus on what the user actually wants rather than trying to artificially deconstruct the fundamentals of design as a concept (which is where The Grid get it wrong, I think).

It's a case of sensible compromise and balance between offering suggestions, taking feedback in frictionless ways, and learning. I've learned from running digital agencies over the years that a lot of the processes that you go through in a web project can be distilled down to heuristic choices and that's what we started with. The rest us refinement through machine learning techniques.

Hey guys, Marc from Firedrop here. I completely understand the negative comments here about The Grid. I was super excited when The Grid came out and paid up the $90 like a lot of people, only to be hugely disappointed when I actually used it. They're pioneers in this space and they've done a lot of really interesting and cool work... but they've misunderstood their users hugely and made too many assumptions about them regarding design decisions. We're taking a different approach where our platform listens and learns more closely from the users rather than whatever broad abstract data set they're using. It'll lead to less "out there" designs that people will be happy with.