Shopify Expert Insights

E-Com Advice from our experienced in-house team

I read the NYTimes piece "How Companies Learn Your Secrets" and was fasincated by people's reactions to Targets data mining practices. In our experience, businesses have the best intentions. They want to provide the best experience for their customers, and they dont want to risk annoying people through irrelevant advertising.

Data mining lets businesses effectively target tailored ads to highly specific audiences. Where people become wary is when they realize that information is being used that they never explicitly told the advertiser. If I give an advertiser information about me, its as if Im opting-in to the advertising and that gives me a sense of control. When I dont know the source of an advertisers information about me, its an eerie feeling. Even if they just extrapolated the information based on people like me, so long as it was an accurate assumption, it will appear to me as an invasion of privacy. The difference between "helpful" and "creepy" in advertising is choice.

One way to not be creepy is to be very transparent about your practices. Targets stone-walling of the NYTimes reporter isnt transparent. If people understood that targeted marketing is about statistical probabilities, and not someone poking around in their garbage cans at night, then they may feel differently. In this case, transparency is the difference between looking like a statistician and looking like a stalker.
Our own lead developer, Paul Reda, recently had success "going viral." On Thursday, he uploaded a timeline biography of 1900's  Chicago private investigator Cora M. Strayer. By Friday it had received over 4,000 unique visitors.

How and why did Paul's Cora Strayer piece go viral? The initial promotion was simple, Paul tweeted to his 145 followers: "So over the [weekend] I researched the life of a woman who was a PI on the South Side in the early 1900s. She was awesome." And then fizzled out. It retweeted only eight times. One of those people did think it interesting enough to submit it to weblog BoingBoing.

That evening, Boing Boing published the article, calling it "a fascinating, and often tragic, timeline of extraordinary adventures." From there it spread to link aggregation site Reddit, and community blog Metafilter. Combined with sporadic blog referrals, it received an additional 2,000 views over the weekend.

It is interesting to note that the article itself was posted as a static webpage, and featured no social or interactive features at all. It was an incredibly simple execution. Its merit was entirely in its written content, not any gimmick.

Paul's Cora Strayer article went viral because it's an interesting story. It's historical truth with some salacious implications, and it had never been told before. People who read it were intrigued, and wanted to share that feeling with others. That's the core of why anything goes viral. It makes us feel something that we want others to feel too.

The inconvenient truth is that there's no way to predict what will go viral, and no way to make something go viral. In this instance, it took unique and engaging content, and a day's worth of work. There is never going to be easy way to do that.
I used to find Quality Scores on Google AdWords very frustrating. They never seemed to make sense. I got so frustrated that I suspended my Google AdWords campaign. When I did this, my AdWords rep called me up and tried to get me to spend my $10k budget again. I refused unless the rep taught me how to achieve high quality scores. Based on that information, and with some experimenting, I came up with a system that always works.

  1. Make a new ad group. Quality scores are historic, you must do this if you have a low score.

  2. Choose one phrase or exact match keyword.

  3. Use that keyword as the first part of your ad headline.

  4. Create a landing page using your keyword as the title.

  5. Wait three days for your click-through rates to be established before adjusting.


You'll start with a quality score of 7 like this. Over time, if you have a good click-through rate, and a low bounce rate, the quality score will go up.