Semantici.st

We recently tried an experiment in selling stuff on the internet. It was a bit of a departure from our normal area of expertise, so it was definitely heavy on the ‘experiment’.

I’ll say it right at the start: it was a complete and unmitigated failure, but not in the way I was expecting at all. Which, in a roundabout way, makes it a success. Just a crap one.

Our plan was pretty simple: take advantage of the current trend for mobile phones built into watches, the availability of easy drop-shipping (which I’ll explain in a bit for those that aren’t familiar with it), and £75 of Google Adwords credit that randomly dropped through our door. Well, the broad outline was simple. The devil, as always, is in the details.

The first thing is catching the trend. I’d noticed the ruinously expensive LG watch phone was released last year. It’s the only 3G watch phone I’ve seen, which probably explains the cost. I’d actually registered a domain for this project last year in May before the LG watch was released, and in retrospect I wish I hadn’t sat on it for a year. Because it meant pushing outside of what I normally do, it got sidelined in favour of things that were less outside my comfort zone.

Watches that are also mobile phones is a pretty niche market, but that’s okay. A niche market is actually what you want for this sort of thing – I’m not exactly going to be able to complete with Carphone Warehouse on price, marketing, or customer service. Google’s information on the number of searches on the appropriate keywords seemed to suggest that it wasn’t too niche. There were enough people looking for these gadgets to support a market, and the bid price on AdWords didn’t seem too high. It all looked good, now to find a product.

Drop-shipping is a way that wholesalers let you sell things without keeping stock yourself. When you get an order from a customer, you place your own order with the wholesaler, and they ship it directly to your customer. There’s a few key things to make it work – they don’t put their own branding on it, so when your customer gets their watch they don’t know that it actually came from someone else. The other big thing is that in order to get a price that actually makes this affordable, you need to look for drop-shipping wholesalers in China. This does mean that the shipping cost is going to be higher – I just absorbed this into the retail price and offered free shipping. There’s another side-effect of this that I’ll come to later.

Working this way you don’t need to have capital to invest in stock, which really reduces the risk involved. The other thing that reduced the risk even further for us was a £75 Google Adwords voucher. We’d always intended that Adwords would be the primary way to get customers, so this was basically a chance for us to try it with no capital investment.

I looked at Google’s keyword tools and they suggested that the cost-per-click (CPC) for the keywords I was after would be around 5-7p. I figured that if I split the £75 over a ten day campaign that would be at least 100 clicks per day. Each sale would have a margin of around £50. Two sales in the 10 day period would be ‘break-even’ point, and would make the experiment a success. That’s a conversion rate of 0.2%. Break-even is considered a success because it allows the experiment to continue, and we could then do A/B testing to refine the site and improve the conversion rate.

A failure would probably mean no sales at all – which is exactly what happened. There’s a lot wrong with the landing page for the ads (which I’m not sharing here, because it’s just frankly too embarrassing), but given the market I was trying to aim for I think it would have worked out okay. It wasn’t terribly sophisticated, but from a sales point of view it was an acceptable start.

In the 10 days the experiment ran we had a total of 82 clicks on the adverts. To sell even one at that volume of traffic we’d need a conversion rate higher than 1%, which would be really amazingly high for this kind of market.

What went wrong? The first thing was that I’d grossly underestimated the cost-per-click for the keywords I’d need. Google’s keyword suggestion tool turns out to be pretty useless and doesn’t tell you what you need to be on the first page of search results, let alone in the top three. A week of obsessive fiddling with bid prices for specific keywords managed to help, but I found that Google still wasn’t displaying our advert often enough. When it was being displayed we were seeing click-through-rates that were all over the place – most keywords were at 0%, some as high as 21%. Even the keywords we had good placing, higher bid prices, and high click-through-rates didn’t get many impressions.

I expected to fail because the landing page would be rubbish, or because the price would be too high, or because people would buy the thing and then return it under distance selling rules, leaving me with a gadget to sell on eBay. I was concerned about the shipping from China resulting in people getting hit with import duty, and the potential trouble that could cause. There was a warning about it in the terms and conditions on the site, but no one reads those.

I didn’t expect to fail because Google didn’t fancy showing my advert enough to let me spend my whole daily budget. After the 10 day experiment I stopped the Adwords campaign, with around two thirds of our credit still left.

From reading I’ve done since then, I think that the longer your campaign’s been running the more Google likes you. If I’d started a year ago the CPC would have been lower (since ‘SWAP’ released their cheaper watch phone this year, there’s been a lot more attention – so a lot more people wanting to get this traffic) and by now it would have been more established. That would have meant investing in the advertising back then, and I don’t think we had the money to risk in this kind of experiment.

Clearly I’ve got a lot to learn about how Google Adwords works. Luckily, when I was at the Scottish Ruby Conference I won £250 worth of books. One of the books I picked up was ‘Google Advertising Tools’. Next time I engage in an internet money-making experiment, I’ll hopefully have a bit more of a clue what I’m doing.

In the meantime, though, I’m out nothing but my time and I think the time spent was a fair trade for what I’ve learned. If nothing else, it’s good to get out of my techy geeky comfort zone and try something very different. I plan to keep doing different things this year – as well as doing my usual stuff too. As the man said, specialization is for insects.