One of the big differences we hope to make with Webtogs, is in the way we display products on the site. There’s no doubt, that the more information a website provides, the better the conversion rate. There is also an emerging school of thought that says you can’t provide too much information.
What do I mean by information? Well, pictures, video, editor style reviews, customer reviews, features and benefits, technical details, etc. Take footwear for example, in addition to showing accurate size information on a pair of boots, it has to be good to show size specific feedback from customers. Zappos and more recently Endless do this, which I’m sure helps customers feel more secure about their impending purchase.
Now, I covered customer reviews in a previous post, so I won’t go on about that here. What I want to do now, is post in 2 parts. Part 1 (this one!) is about product photographs, and the ups and downs we’ve been through, in trying to get this done for webtogs.co.uk. Part 2, to follow shortly, will be about video, and how we see this playing a crucial role in displaying products on the site.
Types of image display
Most sites show a single view of a product, with the vast majority of these allowing a larger version of the same image. Other sites show multi-angle shots of product and even a 360 degree spinable (not sure that’s a word!) view.
Now the single image approach is ok, assuming your product looks almost the same from all angles and has no particular features that require highlighting. An example here would be a football or a frisbee. Just about anything else can benefit from multi angle images.
The 360 degree style displays are interesting. These are normally done on 1 axis, and made by taking multiple photographs and then stitching them together to form a seamless rotating view. To produce a full 3D image, that can be rotated on any axis, requires around 600 photos*, so it’s (a) time intensive and (b) a big ask for the customer to download, even on a broadband connection. Another option, is to model and texture the product in 3D (something we have quite a bit of previous experience with). This requires a fairly hefty investment in software and hardware to do well, but, with a talented operator, can be almost as quick as taking 5-600 images. We took this a stage further and, earlier this year, experimented with modeling a pack in 3D and then exporting it as a 3D SWF object, meaning it would be Vector and play back via the flash plugin. The theory here was we could achieve 90% of the quality, as compared to an image, for 20% of the file size. As it turned out, this simply isn’t true. The files were huge, running to 4-5 MB and the quality was nothing like as good as an image. For the more technically minded, we were using both Maya and Softimage XSI, with eRain’s SWF export plugin. I still believe that this is possible, but in the time we had, we couldn’t make it work.
We remain to be conviced how well the 360 deg approach increases conversions. Some of the stats we’ve looked at, suggest that a large percentage of customers don’t bother to download the image in the first place, and those that do, find it a frustratingly slow process. In our minds, video can achieve a better result for a fraction of the cost (but more on that in part 2).
The other trick used with product images, is the image zoom effect. This normally involves moving your mouse over an image in one view, seeing a real time zoomed version in another view. We call this the ‘gynecologists view’. Whilst we have no data to support our view, we don’t like this approach. The technical merit is obvious, as it allows a very large image to be viewed on a relatively small screen. But, in our minds, there’s something that leaves you feeling less than satisfied. I think it’s because you can see close up detail, but you don’t get an overall large view of the object. It, somehow, seems disjointed. It’s also very interesting to watch older people use these features. They just don’t ‘get’ it in the same way a more net savvy person does. Our test group (all over 55), nearly all wanted to ‘just see a bigger picture’.
So, for now, we are going to show multi-angle photos of our products, with a click through to a larger version. In time, we will add other methods to some product pages, and measure the impact on conversion.
Getting the product to shoot
Well, before you begin to think about setting up a studio and snapping away, you need to get your hands on stuff to photograph. We’ve found this to be a fairly interesting process!. We are now in the position where we have opened trade accounts with around 40 brands for launch. Not one of the brands we are stocking can supply anything more than a single product image, taken from one angle. So, we have to take our own. We spoke with most of the brands and requested sample product to shoot. Now, around 5 of the brands, said “no problem” and sent us gear to photograph. It gets booked to our account and then refunded, when we send it back. Simple. Another few replied and said, “no problem” and then sent us pre-production samples, most of which differed in some way to the actual product we would sell. Not so helpful!, as we can’t shoot these. When you look at the schedules the brands work to, in terms of getting product made, it actually makes sense, but it never occurred to us, that this would happen!
The remaining brands all came back with differing replies. Some simply said “Why do you need samples, just shoot the gear you buy to stock”, well this is fine, except they are (a) new and (b) covered in tags, labels, etc.
Some brands sent us samples of gear, that we hadn’t agreed to stock. Others sent gear that was for next season, so not relevant for our launch, etc. The point I’m making here, is we, naively, thought this would be the easy part. In reality, we were nothing like prepared enough and started the process far too late. It’s very easy to blame the brands for being disorganised here. But actually, this is mainly our fault. Outdoor is not an industry that fully understands or supports Internet retail yet. They also work to very tight deadlines for production and don’t have masses of stock kicking around that they can send out for us to play with and then send back, as they can’t then sell this as new.
Taking the pictures
Now, I’m not a photographer, and have never got involved in product photography until now. Luckily my wife is (although she might not be my wife for much longer when she realises how many products we have to shoot!) and is doing an amazing job with our products.
Here’s a look at the setup we have put together:
That’s Adam in the middle! Below is a standard test shot that we take for every product. The calibration card is then used as reference in Aperture on the Mac, to help get the colour correction accurate.
After colour correction and some editing in Photoshop to remove any background elements and visible parts of Adam our mannequin, we end up with this:
Now, as you might imagine, it takes about 10-15 minutes, in total, to get this right and ready for upload to the website. So, if you multiply 15 minutes, by the number of images on the main webtogs site, it’s quite a time commitment to do your own photography, to say nothing of between 4 and 7 angle shots per product.
Thankfully, our image up-loader is relatively clever (thanks to Dimitar, our resident code ninja), and takes one large image, and scales it down, generating all the various thumbnails and smaller views we need for the site, on the fly. This process is also non-destructive, so we can change the compression settings or output sizes used and re-run the process, without destroying the original, uncompressed image.
We’d be really interested to hear from any other people who’ve been through the same kind of process, for one thing, we still haven’t found a good way of supporting the hoods on jackets. We currently use the ‘Mat Collis Cardboard Conversion Kit’ (patent pending!), which works ok, but is a real hassle to take out in Photoshop afterwards… A large pint of beer is waiting for the answer to this and many other product photography issues!
Tags:
