Tagging past Shopify customers in ConvertKit via API

We’ve recently started using ConvertKit for our email marketing for Stitch People, and I LOVE IT. It’s flexible, easy to create automation sequences (called courses), easy to move people in and out of courses based on any number of events (purchase, link click in an email, etc.), easy to tell exactly which form someone signed up on, and much more.
The first step in my migration to ConvertKit was to start collecting new subscribers and customers into ConvertKit, and to tag them automatically for easy segmentation later. This included writing a small Shopify app to automatically subscribe customers and tag them based on what they purchased. I’ll write another post about this later.

Once all new subscribers were being funneled into ConvertKit instead of Mailchimp, it was time to clean up old data. The first place I started was to tag past Shopify customers based on the product they purchased. We have a relatively small number of products right now (different variants of a book is our main product), but this would work just as well for a large catalog of products and customers.

To do this, I’ll be using the ConvertKit API (documentation found here).

Step 1: Create the tags in ConvertKit

The main API endpoint we’re going to be hitting is the Add Subscriber to Tag endpoint. And since the API doesn’t support creating new tags yet, you’ll need to create these tags through the website first. Head to the Subscribers area, and click ‘Create Tag’ at the bottom of the right sidebar. That’ll bring up the box to create a new tag.

convertkit tag creation

You can come up with your own naming convention for tags, but my general rule of thumb is to be descriptive. You may know what an obscure tag means when you create it right now, but your 6-months-from-now self won’t, and you’ll kick yourself when you have a bunch of people tagged, but you don’t know why.

For example, I use tags like ‘diy-physical-book’ and ‘diy-ebook’ to differentiate between two different variants of the DIY Stitch People book. I prefixed it with ‘diy’ because we’ll be releasing more books later, for which we will have physical and digital copies. For those, I’ll create tags like ‘farm-animals-physical-book’ and ‘zoo-animals-ebook’. These will be easy to decipher down the road.

Step 2: Grab your list of customers from Shopify

Next, head to Shopify and do an export of all your orders. Head to the Orders tab, and hit the Export button at the top. Make sure you select the All Orders radio button, and export the orders, not transaction histories. I use the CSV for Excel, Numbers, etc. option. I’ll eventually be uploading this to Google Drive.

shopify order export

Step 3: Make separate lists of email addresses for each tag

This will likely be the most time-intensive step in this whole process. Now you need to take the CSV from Shopify and filter through several times to generate lists of email address for each tag you want to apply. For me, I created three lists: one for physical book purchasers, one for ebook purchases, and one for physical + ebook purchases. These are the three tags I’m using, and now I need to come up with three separate lists of email addresses–one for each tag.

For this step, I used Google Spreadsheets. You could also use Excel, Numbers, or any other spreadsheet program that lets you quickly and easily filter your table based on column contents. Once you have your CSV file loaded in whatever application you’re using, you need to filter the table by products/variants. In Google Spreadsheets, you can use the Filter feature and then select the ‘lineitem_name’. In my case, I’ve been through several different iterations of the product and variation titles, so for my ebook DIY book buyers, I had to select several values:

filtering lineitem_name

Once filtered, copy those email address and throw them into a new tab in the spreadsheet. Make sure you label the tab so you know which list you have. Repeat this step for any other tags you want to apply.

Step 4: Write the script to tag the subscribers

Now comes the fun part. You can choose whatever language you want, as long as you have a way to make an HTTP POST request. I chose Ruby and used Faraday. Here’s a gist:

First thing you’ll want to do is populate the arrays of email addresses. Take the email address lists from your individual spreadsheet tabs and format them as strings in an array. I used a text editor (Atom) to quickly do this, then just pasted in the result. I was dealing with about 2,000 email addresses total.

Once you’ve created the arrays, head to your ConvertKit account page to find your API key. Put that value in.

Finally, you’ll need to grab your ConvertKit tag ids and populate those into the URL strings. Head to ConvertKit, click on the tag you want, and grab the ID from the URL:

tag url

Put that value in for the XXXX between ‘tags’ and ‘subscribe’ for each tag.

Step 5: Run the script

Now it’s time for the magic. Save the file you’ve been editing and then run it. For Ruby, I just opened a terminal window, navigated to the folder where the file was, and typed ‘ruby filename.rb’ to run it. Mine took about 10 minutes to run to completion. Once that was done, all my subscribers were properly tagged with the product they had purchased, allowing me to create Courses and Segments based on those tagged subscribers.

CSS Framework for Shopify Embedded Apps

I’ve been working on a custom Shopify app for our Stitch People store and was recently looking for a way to make my app look like the rest of the Shopify admin area. Using the Embedded App SDK (EASDK), introduced in February 2014, the app gets loaded within an iframe, sandboxing the whole thing from Shopify’s styles and other assets. After looking around the forums a bit, the only thing I found was a short outdated thread about CSS, with no resolution. It mentions the possibility of a Widget Kit, but I haven’t been able to find anything else on the subject.
I did come across a project on Github–the Shopify Embedded App Frontend Framework by microapps, a Shopify dev shop with a few successful apps already in the store. It has several files and a few directories, but the only file to worry about is the seaff.css file. Include it in your app and then look at docs.html for some examples on how to use the framework. So far so good for me. I haven’t used all the different elements yet, but what I have used has looked great. If you’re building a Shopify app, I’d definitely give it a look.

Shipping a physical product

I never thought I’d be in the business of shipping a physical product. As a freelance developer, I always thought I would only ever offer services, or at most, a digital product like an e-book. So when my wife self-published the DIY Stitch People book, we had to figure out how to get those books from point A to point B.
When we first started selling the book, we didn’t have any sort of system set up. When we needed to ship a book, we’d go to a local place that handles everything from packaging to shipping. We’d buy a padded envelope, stick the book in, seal it, and then head to the front counter and read back the name and address of the customer while the person on the other side of the counter typed it in and generated a label. Then that person stuck the label to the envelope, threw it in a pile with other outgoing mail, and we paid for the envelope and postage.

This process sucked, especially when we had to ship out over 50 pre-orders at once. We were at the store for about 2 hours, and the receipt we got at the end was super long. Here’s a picture of the nice employee holding our receipt:

Stitch People Receipt

On that receipt were the tracking numbers for each package, too. So we went home and typed in each tracking number for each order and sent the customer their shipping confirmation email.

Lizzy and I quickly realized that this process wasn’t going to work. It was time-consuming, expensive, and error-prone. First, we found padded envelopes online at ULINE. Buying in bulk online was definitely cheaper than buying individual envelopes in the store. Problem solved (or so we thought–more on this later).

Next, we looked for a way to generate labels at home. That way, we wouldn’t have to spend so much time at the store reading off the names and addresses of each customer. We love podcasts, so we’d heard a lot about Stamps.com. We signed up and started generating labels at home.

With our new envelopes and the ability to generate our own shipping labels, the next iteration of our shipping process was set. When an order came in, we’d throw a book into a ULINE envelope, go to Stamps.com, copy and paste all the customer information, generate the label, print it out, and tape it to the front of the envelope. Then we’d copy the tracking number and paste it into Shopify, which would send the customer their shipping confirmation email. And because we were using USPS (Media Mail, because we’re shipping books), we could leave the envelopes for our mailman to pick up, so we didn’t have to leave the house to ship. Not a bad improvement.

But this process got cumbersome after a while, too. The copying and pasting from Shopify to Stamps.com wasn’t very slick, and we were still making a few mistakes every now and then when generating labels. We figured that since the customer information was already in Shopify, there had to be a way to easily generate the label and update the order with tracking info without copying and pasting. So I went in search of another solution and landed on ShipStation.

I’ve written about ShipStation (and some of this other stuff) already in another post. It’s a really great solution and is what we currently use. When an order comes in, we go to ShipStation, where I’ve created presets for all the different options for shipping that we use and offer customers (international, media mail, priority, etc). It pulls the order data in from Shopify, we apply a preset, and then generate and print the label. As soon as the label is generated in ShipStation, it sends the tracking number back to Shopify, and the customer is sent their shipping confirmation email. Another big improvement.

When we first started using ShipStation, we were still printing labels on regular paper with our ink printer. It worked, but it was a hassle. The labels we generated were 4×6, not 8.5×11, so we were wasting a lot of paper. And taping labels to the front of the envelope sucked. Lizzy’s mom (who helps with fulfillment) suggested we look at a thermal label printer that worked with ShipStation. We found the Dymo LabelWriter 4XL on Amazon.

DYMO LabelWriter 4XL

With this thing and some cheap generic labels, things got really easy. Now we could process a batch of labels in ShipStation, the tracking numbers would get sent back to Shopify, and we could print on sticky labels without needing ink. The thermal-printed labels (the exact size of the generated label) just peel off and stick right to the package. This was a HUGE improvement for us.

The most recent thing we changed was the envelopes. We were getting complaints from customers that the packages were being mutilated in transit, arriving bent and creased. First we tried contacting USPS about it and learned that there’s nothing they can do. With Media Mail especially, our packages were being tossed around with much bigger and heavier packages. If we wanted to keep using USPS (they’re the cheapest option for us), we’d need to look at new packaging.

I checked out the ULINE catalog and found their Easy Folder Mailer boxes which work perfectly. It’s a tight fit because our book is spiral bound, but it protects the book so much better than the envelopes. It’s a bit more work to put the book into one of these things, but since we started using the boxes, we haven’t had a single complaint about the condition of the book when it arrives. For us, it’s definitely worth it.

And that’s where we are today. When an order comes in, we grab a box, fold it around the book, tape it shut, go to ShipStation, print out the label, slap it on the front, and leave it for the mailman. ShipStation talks to Shopify and sends over the tracking number, which is automatically sent to the customer. It’s not a perfect system, but we’ve found a nice rhythm to our shipping process. And I’m sure we’ll find other ways to improve as the need arises.

Do you ship a physical product? What would you suggest we do differently?

Why does your website need maintenance?

Imagine a single-room log cabin. Like the kind that Abraham Lincoln was born in, with four walls and a roof. There was no plumbing, or temperature control, or electrical wiring. If a log was damaged by dry rot, you disassembled the entire house, replaced the log, and put it all back together again. If the roof leaked, you patched it. Pretty low maintenance.
Now imagine a new house with a tankless water heater, a Nest thermostat, a smart fridge, solar panels, etc. There are a lot of individual pieces, each doing a specific job, and each requiring a different level of maintenance. I spend way more time replacing individual sprinkler heads than I do replacing the flame sensor in my furnace. If I don’t replace a broken sprinkler head, it won’t affect me too much. But if that flame sensor stops working in the middle of winter, I’m going to feel it.

But let’s say you’re part of an HOA and maintenance is taken care of for you. You’re going to pay for that convenience, but at least you don’t have to worry about maintenance anymore. You get the benefits of home ownership without the need to constantly monitor and track all those moving parts that can (and will) break.

Websites are much the same way. Back when websites were just a collection of HTML pages with inline styling and static content, there wasn’t much maintenance required. If the HTML standard changed and browsers started interpreting certain elements differently, you’d have to account for that. But otherwise, you built it once and left it alone.

Today, we’re building websites with a bunch of different moving parts and complex dependencies, each requiring a different level of maintenance. If Facebook changes the way their app authentication works, your Like button stops working. If your WordPress theme relies on a deprecated function, updating WordPress core may cause issues. Or if UPS changes their API, your real-time shipping quotes stop working. Some of these things won’t affect you too much, while others can have a serious impact on your site and business.

And just like there are HOAs for homeowners, you can find website hosting solutions that manage the maintenance for you. Don’t want to deal with WordPress updates? Host your site on a service like WPEngine. It’ll cost more than hosting it on Bluehost or Digital Ocean, but you won’t need to manage and worry about updates yourself.

If a website is part of your business, you can’t expect it to be a “build-it-once-and-leave-it-alone” thing. Chances are, the value you’re providing to your customers relies on a lot of those complex moving parts and integrations with other third-party services that can (and will) break. And generally speaking, the more you pay, the less you have to deal with these kinds of issues. But if you’re trying to run a site on the cheap, you’re going to get your hands dirty and replace a few sprinkler heads every once in a while.

One Year Later

A year ago, I quit my job as a product manager at MoneyDesktop (now MX) to pursue a freelancing career, focusing on WordPress and Ruby development. Through a friend of a friend, I inherited a few spectacular clients that helped jumpstart my new business. Through word of mouth, the occasional Facebook post, and the professional network I had cultivated, I never wanted for new clients–in fact, I was lucky enough to start turning clients down when I realized I didn’t have the capacity to take them all on.
A year later, I’m still independent. About 25% of my time is spent working with some of those same clients. Most of them are at a point where they just need some minor tweaking rather than completely new functionality. I still have great relationships with all my current clients, and I even have a list of some other freelancers that I can pass potential clients on to. It’s perfect.

The other 75% of my time is spent working on Stitch People. If you’ve never heard me talk about Stitch People before, I wrote a post about it back in 2013. In the fall of last year, my wife self-published the first DIY Stitch People Book for people who want to learn to make these kinds of cross stitch portraits themselves. We realized that the business of actually making the custom portraits wasn’t nearly as scalable as producing a book on how to do it, so we focused our attention on the education aspect of Stitch People.

So far, we’ve sold over 1,000 physical copies of the book, and more than 500 digital copies. We’ve shipped books to over 20 countries around the world. Of the 103 reviews on the website, all of them are 5 stars. Needless to say, we’re happy with the success of this first book and are making plans for more books down the road.

My involvement in Stitch People is on the technical and marketing side of things. First and foremost, I make sure the website is working. We’re still with Shopify and probably will be for a while. Recently I looked at moving the store over to WordPress, but it just didn’t make sense. We like the Shopify experience, the mobile apps, and the peace of mind knowing that if our traffic spikes, we’re taken care of.

In addition to just keeping an eye on the site, I’m always running some sort of A/B test through Optimizely. Currently I’m running a test on whether or not calling out ‘Free Shipping in the US’ makes a difference on the book product page. So far, the results are pretty inconclusive, but it’s only been running for a day or so.

I also spend time every day looking at Google Analytics to see if there are areas or aspects of the site that need to be shored up. I look at bounce rates by platform and by referral source. If we’re seeing unusually high bounce rates from a specific source, I’ll dig into what might be causing it and see what I can do to fix it, like setting better expectations upfront through that source.

We’re running ads on several different networks. Facebook is our primary source of conversions, with Twitter and Pinterest contributing a handful every week as well. Twitter and Pinterest aren’t as cost effective as Facebook for us right now, so I’m experimenting with creative to see if there’s a better way to connect specifically to the Twitter and Pinterest crowds.

We’re also running some retargeting ads through AdRoll. Retargeting is a new area for me and I’ve been enjoying the learning process. Again, the biggest thing I’m focusing on here is experimenting with different creative to try to appeal to the retarget audience, both on Facebook and on the web at large. I think this is one area where we’ll eventually see some serious ROI once we get the formula right.

The last big thing we’re focusing on is the community around Stitch People. We have over 4,500 newsletter subscribers and great open and click rates for the industry we’re in. It’s an active and interested community, and we’re coming up with strategies for how to better communicate with them, allow them to communicate with each other, and really start a movement.

So between looking at and tweaking ad performance, coming up with new ad creative, checking analytics, doing research for some of our future plans, and trying to connect with our community, I manage to stay pretty busy. As for Lizzy, she’s the one fielding all the support emails, all the cross stitching, all the designing, and all the one-on-one connecting with community members.

It’s been a good year.