How A Half-Broken Halloween Promotion Smashed Revenue Records (opens in new tab)

I particularly recommend the chapter on virtual assistants. I haven't actually used one yet but that is on the list of things to try.

Awesome - hope you enjoy it. Email me if you have any questions.

I don't know. I appreciate the transparency of it all, but my take away was that his halloween sales smashed revenue despite his promotion efforts, not because of them.

Looking at the promos, it looks like he got 15 (mini site) + 15 (email) sales, which appears to be around $700+ in revenue. Percentage-wise it seems notable, but in absolute numbers, I wonder if the time could be spent better in other ways.

This is super valuable advice. One of the biggest issues I've seen in the technology world is forgetting that we are (often) not the target customer or the largest moneymakers.

There's a kind of minimum barrier of entry to interact on the places we like to hang out (knowledge, interests, etc), so a lot of people tend to get used to everyone being at that level of intelligence/tech savvyness. Then you launch a product that any "tech" person would easily be able to navigate and use, but someone who still mixes up which one is the mouse sometimes might still have some problems with it.

Exactly. And the average web user does not have the aversion that we (techies) do for email. Most people don't use RSS or Twitter so email is their primary medium for online communication.

[bingo cards] is not a huge fraction of my business (4%ish) -- [bingo card creator] (with both brand & generic intent), [bingo card maker], [make your own bingo], [custom bingo cards], yadda yadda yadda, and then I start getting longtail searches from about a thousand activities which cover most of what an elementary school teacher could want in a lesson plan. Obviously, most of them don't contribute a sale in any particular month -- heck, 80% of them have yet to contribute a sale ever -- but the aggregate economics work out to staggeringly efficient. This is one of my Big Ideas, see the greatest hits section on my blog for it done to death several times over.

Every reason people have for buying bingo cards in September is a reason to buy bingo cards in October, plus the various promotional activities, plus my AdWord spend goes vertical after around October 15th.

Normally, my AdWords performance is constrained by inventory: I could literally buy every click available for sale from Google for the keywords I care about at prices which make sense to me. When I don't bork the heck out of my conversion codes, I'm also inventory-constrained during Halloween, but inventory levels are much higher than usual.

BCC doesn't have a staging environment, owing largely to being cobbled together with bubblegum from back when HTML written in Notepad was cutting edge web technology for me. I also use a lot less testing than is considered ideal in the Rails community. Getting a staging environment is probably worth doing, but the level of coverage with Selenium tests I would need to catch things as obscure as "had an image delivered over HTTP on the landing page" strikes me as not likely to have the highest ROI of all things I could be doing.

(I've been interviewed about what I do for reliability before. The short version is checklists and automated warnings when stuff goes wrong. I've set up some automated things to catch issues spiritually similar to the ones described in this post already.)

I think that's the same conclusion Patrick reaches in his postmortem.

I hope so too. Regarding halloween though, I'd wager that people who bought Minecraft over the weekend probably did it because of word-of-mouth excitement over the halloween update (in other words, they bought it because they heard about it from other people who were excited about the halloween update, and not because they were waiting for the halloween update themselves).