Eligibility Questions
I work with an agency on repurchase flows today already, will this replace them?
No, we are fully willing to collaborate closely with your agency's current efforts. Black Crow's Predictive Repurchase product revolves around three essential elements: timing, dynamic content, and an optimized storefront. We are a technology service provider and not a marketing agency so we highly appreciate their assistance in creating email templates that can effectively enhance the Click Rates on these messages.
What subscription providers do you support for the Predictive Subscribe use case?
Recharge, Shopify, Stay.ai/retentionx, Loop, Skio and Yotpo.
Please let us know if you use any other provider for subscriptions and we can test and validate if our solution will work with that provider.
Event Timing Questions
How do you know when someone runs out of something to trigger a replenishment event?
Our model predicts the best time to send someone a reminder email to replenish a product by looking at a brand's Shopify data patterns for how often and on what cadence the particular user and all users tend to reorder the product.
It cannot know for sure if someone has run out of the item or not but you don't necessarily need to wait until someone runs out of something to remind them to reorder. Users often buy things again when they are ALMOST out of the item rather than when they've completely run out.
If a customer is eligible for multiple repurchase flows on the same day, how do we determine whether to send a replenishment, cross-sell or subscription email trigger - and prevent multiple sends?
No user will ever get more than 1 Predictive Repurchase event on the same day so if multiple are predicted on the same day, only 1 event will be chosen and sent (the event that we predict will maximize your revenue).
We already have manual triggers for replenishment, cross sell, or subscription flows. Why would these AI triggers perform better?
With manual rules, brands have to pick a hardcoded number of days after an order before sending a repurchase email, but we know that every user's duration between purchases is different. With Black Crow AI's event timing, you don't have to treat every user like they're the average when it comes to sending repurchase emails. Our AI predicts the optimal time to send each individual a repurchase email based on their own historical data and real-time behavior.
This leads to higher flow revenue than flows with hardcoded durations. It’s also typically significantly easier to manage this flow without lengthy manual rules and decision trees that would need to be managed and updated over time as user behavior changes throughout the year.
Email/SMS Flow/Journey Questions
There's some wording in the flows you set up for me that I don't like. Will you change it for me?
Black Crow AI sets up flows using a default template for message wording. You can use the default wording out of the box and achieve great incremental revenue results. However, if you want any changes made to flow message wording or styling after set up, Black Crow AI doesn't make those changes for you but you are welcome to change the wording and styling of the messages however you like.
Why do you have a plain text email rather than a styled/branded email with graphics?
When sending an email to a previous customer, incorporating a personal touch is typically beneficial as it resonates well with customers in this particular flow. Our team has conducted consistent A/B testing, which has clearly demonstrated the effectiveness of this approach, resulting in click rates that are 50-80% higher than HTML-based styled/branded emails. If you still want to make styled emails, you are welcome to change the wording and styling of the messages however you like.
Can we change the emails to include styles/graphics/branding/etc?
Yes. It's not necessary to make those changes to see positive revenue impact from the flows. We've seen great results with text based emails. But contents of these email and SMS messages is ultimately something you can change and that you control.
Your email set up is HTML. If I switch to drag and drop, it removes the data. What's the best way to recreate it in drag and drop?
If you want to switch to drag and drop email editing in these flows and are worried about losing the details of what variable names are used in the email and links, you can clone the flow that Black Crow made for you so you have our original set up that you can use as a reference. Then, in the cloned flow, you can change to a drag and drop layout and easily be able to copy and paste wording, variables and link URLs from the version Black Crow initially made to preserve the variables and link functionality.
Can we A/B test discounts and messaging in repurchase flows?
Yes. You will just need to duplicate the emails into an A/B test and update two things:
- The copy in the emails to include the new percentages
- The deep links in the emails to include the new discount code on the URL
You will manage these A/B tests rather than Black Crow AI managing them. Your contact or documentation from Klaviyo, Attentive, Postscript or Yotpo can support you in how to set these up.
Someone got a Predictive Repurchase message before their first order was fulfilled (or very shortly after). Can we prevent that case?
In Klaviyo,
there a couple approaches to address this. You can use 1 or both of these approaches based on your preference.
Approach 1:
If you know roughly how long it usually takes for orders to be fulfilled, you can add a profile filter to the relevant flows like "Placed Order zero times in the last 7 days". That would work well if you know you fulfill orders in about 7 days to make sure anyone who has very recently bought something won't get a repurchase message.
Approach 2:
You can check Klaviyo's Fulfilled Order metric and only send to people who have had at least one order fulfilled. Adding this as a profile filter in any flow where you want to have this filter would ensure that only people who have had at least one order fulfilled in the past would be contacted.
Note that approach 2 could still result in messages sending for customers who have a fulfilled order but now have another order that's unfulfilled.
However there's no requirement not to send to someone just because their most recent order hasn't been fulfilled yet.
In Attentive,
there's no fulfilled order event to use so Approach 2 won't work there.
But a brand can use Approach 1 there like this to branch on whether a user has made a purchase in the last X days (e.g. 7) and exit the journey before receiving any message if they have.
Product Recommendation Questions
Are your product recommendations just best-sellers or are they personalized based on a data model?
Black Crow's product recommendations are not just sending a list of your best selling products. They are personalized based on a data model to get the products that we predict the specific user receiving them is most likely to purchase.
How can I ensure you don't recommend a free product, something I don't sell online, or something that I just want to deny list?
Black Crow automatically ensures that each item that gets recommended is in stock, has a price above $0, can be sold online and, in the case of Predictive Subscribe, is available for subscription. If there are any specific products that do not meet these criteria and you do not want us to showcase them to your existing customers, please tag those items directly in the Shopify Admin UI with the following tags:
- black_crow_do_not_replenish
- black_crow_do_not_subscribe
- black_crow_do_not_cross_sell
black_crow_do_not_replenish will prevent the product from being included in Predictive Replenishment messages asking users to buy them again.
black_crow_do_not_subscribe will prevent the product from being included in Predictive Subscribe messages asking users to subscribe to automatically replenish these on a schedule.
black_crow_do_not_cross_sell will prevent the product from being included in product recommendations for Predictive Cross Sell messages.
For Predictive Cross-Sell, can I manually suggest products for a user to see?
While it's technically possible to put whatever content and links you want into Predictive Cross Sell messages, we recommend including our product recommendations that are generated with machine learning. Our system relies on a sophisticated neural network to generate personalized product recommendations. This approach takes into account each customer's unique history, which includes past purchases, browsing patterns, and interactions through emails or SMS. We have found that individual buying behaviors can vary greatly, meaning that what may work as a sequential recommendation for one person may not be applicable to another.
Our automated system excels in detecting and adapting to emerging trends, ensuring that the recommendations provided remain relevant and tailored to each user. While we appreciate the thoughtfulness behind manual suggestions, our model is designed to autonomously identify the most relevant products without the need for manual input.
Our system leverages continuous learning to stay up to date with dynamic changes in the market and consumer behavior. By utilizing this method, we can guarantee that all recommendations are data-driven and personalized.
How do product recommendations handle seasonality?
Our model stays up-to-date by being regularly retrained with the latest brand order data, capturing the most recent purchase trends
- As buyers begin purchasing new products, the model learns to adapt to their new interests
- This continuous training occurs on a regular cadence, ensuring that recommendations are always fresh and incorporate the latest purchase trends
- Finally, recommendations are assembled just before email delivery to guarantee they're current and reflect the most recent model updates and purchase trends
Why are you recommending products where I added the appropriate exclude tags?
If you've added these tags to a product they will not be recommended going forward:
- black_crow_do_not_replenish
- black_crow_do_not_subscribe
- black_crow_do_not_cross_sell
Note that there's an up to 24 hour delay after these tags are added and when Black Crow AI will stop recommending them on new events we send.
Also note that only events going forward after you add those tags to a product will exclude the product. All events sent in the past (before you added the tags) and up to 24 hours after the change may still include the product in recommendations.
Storefront UI Questions
Why can’t users find products outside of the previously purchased and recommended products on the Black Crow AI storefront?
This intentional design choice of the storefront aims to simplify and expedite the repurchase process. Users can easily navigate to the homepage by clicking on the brand logo in the header, which is a common functionality found on most websites. Should users desire to explore the shop's products once more, they can always visit the main website through the logo click. However, our current flow is specifically optimized to cater to those seeking a streamlined repurchase experience.
Can our own developer make updates to the Black Crow AI storefront UI?
Your developer cannot make updates to the storefront UI but you can request for us to make style updates to colors and the logo used on the storefront any time which we will act on promptly.
If you have layout suggestions in mind, we are interested in hearing your ideas and feedback for consideration for our product roadmap but cannot guarantee we will act on all requests or any timeline to do so.
We have many customizable products, can we use Black Crow AI’s personalized storefront?
If a user needs to do more than just select a variant to buy one of your products, i.e. if they need to customize the product before ordering (e.g. if they need to upload a photo or need to type in a custom label to buy a product), you should not direct them to Black Crow AI's personalized storefront as it doesn't support this customization.
Instead, you should change the links in the Predictive Repurchase flows to redirect to the most appropriate place for your brand so you can still take advantage of the Predictive Repurchase product for the automated event timing predictions.
Can I use Predictive Repurchase even if I don’t want to redirect users to their personalized storefront developed by Black Crow AI?
Yes. Predictive Repurchase can still be a useful tool even if you don't want to use the personalized storefront component of it. The automatic event timing component can still help you generate more revenue even without pairing it with the storefront.
What do I need to change in Klaviyo to have Black Crow AI Predictive Repurchase flows redirect to my product detail pages instead of Black Crow AI’s storefront?
What do I need to change in Attentive to have Black Crow AI Predictive Repurchase flows redirect to my product detail pages instead of Black Crow AI’s storefront?
What do I need to change in Postscript to have Black Crow AI Predictive Repurchase flows redirect to my product detail pages instead of Black Crow AI’s storefront?
What do I need to change in Yotpo to have Black Crow AI Predictive Repurchase flows redirect to my product detail pages instead of Black Crow AI’s storefront?