Show HN: Auto Swatch - New-car research that doesn't suck

by nicksergeant • 15 years ago

15 40 15 years ago

The site is pretty, but I would hardly call seeing what different cars look like in different colors research. I guess if that's what's guiding your purchase, it makes sense, but for me, appearance is nowhere near the top of the list.

nicksergeant15• 15 years ago

When my wife and I were car shopping, I noticed an emotional pattern. She cared about color, style, etc. My target audience is women 25-40.

DanielStraight3• 15 years ago

If the goal is to shop by color/style, I would filter out the cars that aren't available in the selected color. Showing the little gray boxes provides no benefit at all. I would also eliminate the pagination on the listings. Pagination almost always detracts from the usability of a site. In your case, it makes it much harder to compare vehicles within the search results. And, when you remove the pointless grayed-out boxes, the page will never be that long.

nicksergeant15• 15 years ago

Also, I'm working on acquiring the actual data set needed to turn it into a viable research platform. It's coming. Costs money, takes time to arrange deals with providers.

randallsquared• 15 years ago

Just as a counterpoint, for me, cost and appearance were the most important aspects. Efficiency, safety, and power are all good enough in new cars that I don't much care about the differences. Storage space was also a consideration, but it's closely related to appearance, of course.

naner• 15 years ago

Still actually kind of sucks. There's not enough filters.

Also they seem to have conflated "Sportscar" with "Convertible" moving all high performance hardtops under "Sedan/Coupe". This provides the odd situation where an Anston Martin V8 Vantage shows up in the same results as a SMART car.

nicksergeant15• 15 years ago

Yes, there are some category mis-matches. It's a work in progress.

CWIZO• 15 years ago

Categorization of cars is messed up. There's a BMW 5 series in the hatchback section. And a VW Beatle in the sports section ... and countless other examples.

Is "seeing cars in different colours" the only feature you are offering? If you are planning on adding more comparison features then I'd wait with "launching" as this is mostly useless and user's wont remember you. Pretty site tho :)

nicksergeant15• 15 years ago

We'll have all data on each vehicles soon. Something similar to this: http://twitpic.com/4mm384

The categorization is messed up, yes, I'm working on a flabby set of data for that. That'll be fixed soon, too.

andymoe• 15 years ago

It's a really great start and I have been thinking a lot about this problem as every year during ski season I pine for an AWD vehicle of some kind and start bouncing between car manufacturers sites. I search both visually but I also like to hit google for a specific feature like AWD or TDI etc and use that as a starting point. If you sprinkle in a few feature details - even if it's meta only for the search box - then I think it will become even more awesome. (And probably my starting point for all my car research)

Finally, I realize this is probably a first version but once I find a car I like you should make an effort to send me to the the manufacturer site or even a simple google qry that will find what I'm looking for.

I think your site is about 100 times more sticky than carwoo since right away I am looking for cars - if later I want dealers to compete for me that's great. I would seriously consider that you borrow their business model or that they borrow your interface idea.

d5dhatch• 15 years ago

I'll admit I'm in the category of color-first car shoppers. Well-done and quite beautiful. Outside of color and type I might add in price next and mpg. Then hand the search off to another engine to let folks fine-tune criteria. Don't get too bogged down in needing all the data.

nicksergeant15• 15 years ago

My plans exactly. Stay tuned!

robin_reala• 15 years ago

US-only data at the moment. Internationalisation plans?

nicksergeant15• 15 years ago

Plans, yes. We've got http://autoswatch.co.uk and would love to hop into that market. We've got our hands full with US at the moment, though.

Killah911• 15 years ago

It's very interesting. It's a work in progress, but the canvas looks good. One of the huge problems with buying a new car is the sheer amount of "comparison data" available. It's near impossilbe for someone to make a good decision (good according to "How we Decide" & "Happiness Hypothesis") with more than 5-6 variables. I like the fact that it only presents one very simple dimension of choice so far. Now, if you can figure out a way to not overwhelm a buyer, and keep the choices low, by eliminating vehicles of the wrong color etc from a search, it really would be new-car research that doesn't suck :)

vaksel• 15 years ago

nice visually, but kinda crappy functionally.

that said I did find the ability to switch between models useful, since it lets you see just what gets changed visually between the C63 AMG and the regular C-Class, the M3 and 3-series, Base corvette and the Z06 etc.

That said for some reason you haven't applied this to Audi(i.e. A4/S4)

pauldisneyiv• 15 years ago

I run a startup providing inventory services to automotive dealerships and make use of the eVox product. Great images.

That being said I would be very careful how your licensing contract is worded with them. You don't want to get burned and most everything in our space is charged on a dealership by dealership basis.

The idea itself is interesting and I like anything that disrupts our industry. Keep working on it and feel free to reach out if you'd like.

nicksergeant15• 15 years ago

We were in talks with Evox for a long time, our contract very specifically allows advertising-backed public display of the photos. They've seen the site and love it. FWIW, the folks at Evox were awesome to deal with.

vlucas2• 15 years ago

Those photos are gorgeous. I made http://autoridge.com and I had a lot of trouble finding good vehicle photos. I ended up using the Flickr API to pull in photos for both the website and the mobile app. Most of the resulting pictures are good, some are seemingly random or mis-labeled, but none of them are anywhere near the quality of what you have there.

kmfrk4• 15 years ago

Just in case you miss it, pauldisneyiv's pointed out that they're by EvoX: http://www.evoximages.com.

nicksergeant15• 15 years ago

My original plan was to scrape photos off Flickr / MFR websites, but I quickly realized that's not even close to legal :). Evox does a great job.

vlucas2• 15 years ago

Scraping photos is a lot different than scraping data, I fully agree.

But for the pictures, I am using Flickr's API, and it's not illegal. I have read the terms carefully, and since the site it not a commercial one, I am well within the public API terms. I also link to and credit the source on every photo, as I know each individual has the rights to their own photos.

kmfrk4• 15 years ago

UPDATED: Expanded on the categorization aspect.

Your design is absolutely gorgeous, and that's worth visiting the site for alone.

The categorization is a little wonky, and while it may technically be true (if it is), I think it's going to confuse other people than me; Sedan, Hatchback and Crossover seem and look to similar judging by their icon. Maybe they should be merged, or the main categories should be based on something else. To someone who knows jack about cars, they all look like station cars to me. Heck, upon closer inspection, some of the Hatchbacks and Crossovers look like SUVs or Minivans.

Some people will probably be interested in several similar categories. I'd try enlisting the feedback of your users to understand how they categorize cars and apply at least some of that methodology.

Think about how people buy cars; if you need a big car to transport heavy loads or compensate for something, you have an SUV.

If you're a mother of three, you'd want a "Family Car" (although that's a tough sell to the husband).

Some people, myself and probably some women in their 20s, like "Small Cars", like Mini Coopers, Fiat Grande Punto and their ilk. The Punto is a great city car that's cheap on petrol, easy to navigate, and even easier to park. These cars are like the purple Dell laptops to some women.

Then there's the Sportscar (or Mid-Life-Crisis Car, but let's just call it a Sportscar). No need to delineate convertibles nor two-seaters and so forth.

I'm sure there's a general consumer pattern - a car buyer archetype - that you can use to your advantage. Quora has so many quirky people that there might be someone with that knowledge. This is probably a vital thing to figure out, and basing it on buyer trends might be a good idea - as long as you don't create category frameworks that overlap in domains.

As for relevant metrics, I would suggest MPG (bio/electric/hydrogen vs. diesel vs. unleaded) and the Euro NCAP[1] score. The latter probably requires API access, but nevertheless.

EDIT: What are those coloured boxes supposed to signify? Granted, I'm an idiot, but I thought they were filter links for age groups at first.

[1] http://www.euroncap.com.

kmfrk4• 15 years ago

Another thought I just had: approach your design like you were an online car salesmen - not the Simplified Wikipedia of cars - or that you earned an affiliate commission on every car "sold" through your site.

That way, you would be more focused on better conversion rather than exhaustive classification, and have a completely different incentive. This would affect the categorization and user experience significantly.

nicksergeant15• 15 years ago

Agreed. I think this will become more apparent once the functionality of the site is fleshed-out. The goal is to get people's asses in (the right) cars, not just tool around with the site.

duck• 15 years ago

Your search doesn't appear to handle some characters: http://autoswatch.com/search/?q=F-150

gregpilling• 15 years ago

The search couldn't find the truck as 'f150' either. I have an automotive site (retail ecommerce) and it is amazing the variations on truck names that people enter to find what they want. You could probably get some good search terms to alias if you did some research with the Google Adwords keyword tool. You could make F-150=F150=F series=etc...

nicksergeant15• 15 years ago

Yeah, right now there's some issues with the dataset I've got. Dashes not being dashes, etc.

3pt14159• 15 years ago

Take out the Google ads. If your goal is to make money then there are ways to do that, but don't throw up ads until you're out of ideas.

grandalf• 15 years ago

I have a few comments:

- nice design

- why use categories? I tried to find my car in two categories and couldn't find it. I finally gave up and searched, and still didn't find it (no fuzzy matches).

- Why not use a flip approach to flip through the car detail pages and star ones that you want to do further research on? Starred items could then be viewable as a features/price matrix.

- What is the problem that the site is solving for people? In my opinion it's "What cars are out there, what do they look like, and how do they compare?".

- I think that info about cars that are currently available with big incentives or other discounts would be useful. Two seemingly smilar vehicles could have a very different monthly payment b/c of incentives and as a consumer I want to know.

PonyGumbo• 15 years ago

This is neat, but I think it would be much more useful if you could set a ceiling on price before displaying results, or have the ability to sort by MSRP in individual categories.

nicksergeant15• 15 years ago

It's coming. Gotta get the data first.

hnsmurf• 15 years ago

I wouldn't use this honestly. I don't get anything from it that I can't from manufacturers' sites. Every manufacturer has a "build your own car" link that has colors you can choose. Theirs often even let you choose interior colors too.

Manufacturers' sites suck, but this doesn't give me the info I need to skip them, so I'd now have to use two sites instead of one. This site is just a subset of the manufacturer site I'm already stuck using, and to be honest their color pickers work better.

hkuo• 15 years ago

Having worked one of the top auto manufacturer websites from top to bottom, MPG and Seating are the next two things you need to add. Particularly right now, MPG is top of mind as gas prices are skyrocketing. And seating is always important. It helps to know whether you'll be able to fit your entire family.

Next comes trunk space and safety ratings. Can I fit what I need to given my particular lifestyle? Will I have a better chance of survival in a crash?

Awards help as well, such as Motor Trend or J.D. Powers. Those factor well into the decision-making.

Funny enough, random things like number of cup holders is a deciding factor some people. Will there be enough cup holders for everyone's drinks?

Regarding the interface, I think it would be particularly awesome if filtering and display updates worked using this jQuery isotope plugin: http://isotope.metafizzy.co/demos/filtering.html

nicksergeant15• 15 years ago

Good tips! There is a huge opportunity in this industry on the web. Current car research sites have no human touch - things like cup holders and basic safety ratings. The isotope plugin is pretty neat, but I'm not sure if it's worth spending time on at this stage. It'd be a neat easter egg to introduce sometime, though ;)

kmfrk4• 15 years ago

I just clicked your "Subscribe", and it turned out to be a link to a newsletter subscription. Nowadays, I equate subscriptions to RSS, so I would rename it "Newsletter" instead, if you want to keep it down to one word.

nicksergeant15• 15 years ago

Just an update - added MPG, MSRP and MFR site data on Auto Swatch: http://autoswatch.com/sedan-coupe/

Show HN: Auto Swatch - New-car research that doesn't suck

Related Stories