The eBuyer sale cock up

Posted on 28 November 2011 by Pete Duncanson Tweet

The online computer store eBuyer launched a £1 sale this morning, since then the site has barely been available due to being overloaded with requests. Here we talk about where we think they might have gone wrong and what they could have done differently.

Launching a £1 sale is a bold idea. It is bound to generate sales and possibly more valuable bring in new customers and expose them to the eBuyer brand (which I have to say I'm a big fan of most of the time). I think the issue is probably over ambitious marketing department and under supported web team.

Right idea, wrong implementation

Drumming up new business

If this promotion worked then everyone would have been talking about it, similar to the old days of budget airlines; "fly to ireland for a £1". Everyone would also get to experience eBuyers awesome delivery service too. I don't know how they do it but on numerous occasions I've ordered something late at night to be woken up early the next morning with it at my door. Its magic assuming you can get online to order it in the first place.

To get it right though you actually have to be able to deliver on that promise. You have to actually get to Ireland, you actually have to be able to buy that product for a £1. Having the web site fall to its knees due to the traffic generated makes this impossible. Yes you are bringing new customers in, but then instantly frustrating them. No way to build repeat business. Budget airlines get away with making mistakes some times as they only do budget flights, you know you are getting what you pay for ("it was a steal so I can't expect it to 100% perfect"). EBuyer though offer 1000's of products at normal prices all year around, new customers won't see any of this though due to the site not even responding.

Everyone IS talking about eBuyer, the trouble is that its not in a nice light (just check out their Facebook page). I'm not a believer of Mr Bransons matra that all publicity is good publicity, right now they are getting a hammering.

Reward repeat customers

Those of us who shop with eBuyer regularly (like us) might have signed up to their Facebook page (40,000+ followers) or followed them on Twitter. Ebuyer where smart enough to want to reward these "hot customers" by giving early access to the sale. This was a smart move but even this was too much traffic for the site to handle. Not much of a reward that.

When the site did come back up (its been down more than up since the sale started) marketing instantly updated the Facebook pages to announce it which again brough 1000's of requests down on to the servers in an instant. So far they have done this several times and each time killed the servers, which again is nothing but frustrating for users.

A better solution might have been to realise that some users would keep trying regardless (like me, it was busy so I've been checking in every hour or so regardless of their updates). If they had kept quiet about the site being back up they would at least get some natural traffic (caused by residual from the hype of the sale) which would be of benefit at least.

Part of the problem is that trying to get these offers out to everyone means that on the whole even if the site does come back there is not enough stock to satisfy demand so 1000's of people are just getting disappointed and left feeling robbed, no one likes giving up their time on fruitless promises of bargains to be had.

Regular customers are also penalised by the fact they can't get on the site at all, no way to follow up existing orders, print out invoices, start a return or buy that urgent bit of kit. This is also no way to reward loyal customers for the sake of a quick gain with a bargain sale.

Clear out old stock, make way for new

Low prices should get any old stock shifted quickly, £1 prices should see it fly off the shelves. Stock sat on shelving is dead money, it takes up room that might be better used for products that sell better. This was a great way to clear the decks. With the site down though even this aim is not met. Even when users are managing to buy a product, due to an over sight in the ordering process toomany orders are getting through than there is stock. Lots of users are reporting getting "sorry that item it out of stock" emails after they have purchased their items. This does a lot of damage to the trust worthyness of eBuyer, it looks like some slight of hand when in reality the order process might not be doing a simple stock level check at the point of charging the card. Again I'm assuming the site was never build to expect this much demand for single items of stock.

Get some valuable up sale revenue

Ryan Air used to sell what was jokingly called the most expensive cheese sandwich in the world, I believe it was about £7. You know what though, if you got a ticket for £1 you didn't mind going crazy and having a blow out on the plane by buying a sandwich, crisps and a can of coke for £14. It was still a bargain considering. The same could have worked here, users coming in might have seen other items they liked (or experience the service and become a repeat customer) but again without getting online they can't do that. Ebuyer has good prices, its not all basement bargain stock they could have introduces a whole new load of customers to what they can offer.

What they could/should have done

A sale this big can't have just been implemented over night so several departments within eBuyer should have been involved.

Planning for the un-plannable, how many gritters do we need?

I hope someone asked the website team if the site could take the increased hit. Its a hard question to answer though:

Marketing: "If we do a promotion can the site take it?"
Web Guys: "Sure its handled your newletter promotions in the past no problem and we have about double the server capacity to spare"

Marketing guys then go and launch a campaign bigger than anything else before and spread the word via the numerous channels they have all at once. At the stroke of 11pm the site opens and crashes. Hard for the Web Guys to plan for that one. Lets say the Web Guys where super cautious and said they would need 4 weeks to prepare and 8 new servers. Can't see that one being signed off either given my imagined first response above.

We had the same problem here in the UK regarding the recent years of heavy snow. How do you predict how much snow you are going to have, you can't so you have to stock pile massives of grit in advance which you might not even use year on year. After 10 years of light winters questions get asked about the gritter budget each year and it gets trimmed back to a "more realistic" level. Then we have two years of super heavy winters, run out of grit, crash cars and the country grinds to a halt losing more money through lost production in a day that would have been lost keeping the gritters going for years.

Selling clouds

It's at this point that Cloud computing will be bouncing around your head. You should have built it to scale, in the cloud. Its a lovely stock answer. But when you've built a site up like eBuyer over years it's hard to make the switch from solid servers you know and love to "new fangled cloud computing".

The website would need some major modifications to get running in the cloud, it not just a simple job of uploading it and you're done. You need to plan for it and build it in to the code itself for it to work well or you end up spending more time fighting with the cloud than working with it, have you ever tried punching clouds? Its a losing battle.

So technically yes it could have been in the cloud, heres the problem though, how do you sell that to management. How do you tell them that they need to invest 3 months of work and recurring monthly costs (despite having paid upfront for new servers only last year...etc.) to get the site running "in the cloud" on the off chance there will be the equivilant of a hard winter.

Our local council had the same problem, my gritter (we have personal ones up in the North, he runs the scrap yard at the top of my hill) had been gritting all our roads for year with the same old trucks, the same ones his dad used before him. He wanted to invest in new ones but to do it needed a new 10 year contract to make it worth while. The council refused as they felt safe after 16 years of mild winters. So he scrapped his fleet and no longer grits. That year we had 6 weeks of snow and 18 inches fell in a day. My car did not move for 3 weeks.

For the first time none of our local roads where gritted. A freak event I hear you cry. The following year the same again. My local gritter now has a brand new fleet of truck parked up at his scrap yard, I believe someone finally saw sense and I look forward to driving safely again locally. You can't sell that sort of change easily to management, the numbers don't stack up, yet what they lost for what it would have cost hurt them a whole lot more and I'm pretty sure my gritters rates are now higher than they would have been had they just renewed.

Throttling features

During the busy time for some of our clients we have to switch off (or throttle back) some of the features and functionality on their sites. We can lose some of the fluff to ensure that the core stuff gets through to the client to allow them to do what they came for, buy products. If these features a built in from the off then it is easy to throttle back a website to make it lighter, leaner and therefore faster to server.

Removing some features from the site can drastically reduce the number of Ajax requests, CSS/JS files, Database calls, etc. that a server has to do which means it can handle more load which means it can stay up and responsive when in high demand. Really smart website can even have this built in so the Web Team don't even need to be on call all night to monitor and manually edit config files.

Some concert ticket sites do this really well, sending a simple stripped back page to users without all the bells and whistles when large events are annouced (the latest Stone Roses ticket sales was the last time I saw this and it was a nice touch, I still didn't get a ticket though...)

Here though the techies often have to go head long against management once more. Rather than see that it is better that 100% of users get 80% of the sites functionality and still at least book, management will want 100% of them to get 100% of the functionality.

Logic being that management invested in those features to sell more products, this offer is intended to get up sell and to do that they need all those features, thats what they where built for! They then fail to realised that these constraints mean that no one gets to see any of the products, the site is on its knees. Another tough one to sell to management until of course it happens on their site.

Techies seen as the bottle neck

Techies, true techies who love doing their job love to solve problems, problems that management don't even know exist. Trouble is they tend to suck are explaining to management what those problems are. How can you sell a solution to a problem that the buyer does not understand, you can't get that buy in. Without it you can't get the solution in place. Without that you are flying without an ejector seat, I can imagine the cigar chewing exec now "Our planes don't crash. Do you know how much these seats weigh? Think of cost of the additional fuel we will need over the life time of the plane carting around this luxury bit of kit we'll never use!"

Techies need to get better at defining problems, to do that they need to attach £'s saved/gained to them. Management need to listen to techies more and give them some grace that some of what they have to fix is really hard to explain and should be taken on merit.

Techies like solving problems, to do that they sometimes have to do things that don't make sense to management. That killer new feature will have "to wait until verion 8 is released next month as we'll have the new database layer in place then". They don't do that to spite management, they do it to protect the site as a whole and keep the core function of the site running regardless of the latest must have whims/ideas of the day. Techies are in the tenches long term, winning the work not just the battle, that what you are paying them for so listen to them and trust them.

If they say they need it then management really should listen, it can save a whole lot of pain the long run. They tend not to lie too much. Besides, one day it might be you who needs that ejector seat...