Myself (Grace Kim)
Fellow Designer (Shaunte Thompson)
May 2021 - July 2021
Lead Designer
By the end of 2021, more than 70% of American households (93 million people) have shopped for online groceries during the year¹. The online grocery industry has seen meteoric growth and estimates show that online grocery will continue to climb, accounting for 21.5% of total U.S. grocery sales, or $250 billion of the $1.16 trillion grocery market, by 2025². As a prominent player in the industry, Instacart holds about 25% of the online grocery market share³.
With Instacart’s enormous growth came the increase in media coverage of protests by Instacart Shoppers over the harsh penalties of the rating system. Numerous accounts voice how Shoppers are punished as a consequence of customers’ low utilization of the out-of-stock-item refund or replacement system.
.png)
…we've done a boycott, and we've done a protest, and we've done a walk off, we've asked customers to delete the app.
Why Do Gig Workers Want You To Delete Instacart? by Errol Schweizer
Zahara went from a five-star rating to a 4.94, and with that, her weekly earnings plummeted. Though she worked the same number of hours, Zahara’s weekly income fell roughly $650 a week
Unfair ratings cost some Instacart shoppers hundreds a week. Here’s what’s happening by Johana Bhuiyan
Shoppers are also subject to low ratings when they get orders wrong, usually because a store was out of a certain item, which in turn keep them from higher-paying orders.
Why Be Mad at Your Instacart Shopper When You Could Be Mad at Instacart? by Jaya Saxena
Shoppers should not have to live in financial and mental paranoia that one or two customers will demolish their income, livelihood, and family security with the swipe of a finger.
Instacart’s harsh ratings system hurts grocery delivery people like me by Ehud Sopher
If the customer doesn't respond when you need to replace the item ... you can't sit in a store for an hour waiting for a customer to respond. And some customers are upset with that ... so then they'll give you a lower rating.
Instacart shoppers challenge ratings system by Faith Abubey and Erica Y King
Instacart’s inability to handle unfair ratings from customers – ratings motivated by supply issues and other elements outside a shopper’s control – are also hurting workers.
‘Tired of being stepped on’: Instacart workers urge customers to delete app by Gloria Oladipo
"Customers often get very, very angry with shoppers because we've had to refund an item or maybe we bought the wrong item, because Instacart will prompt shoppers to choose a replacement," Jen said.
An Instacart shopper reveals the 4 biggest mistakes customers make when ordering groceries online by Melissa Wiley
When Customers don’t specify refund or replacement instructions and aren’t available for communication about an item, they end up receiving an unwanted refund or replacement by Shoppers. As a result, Shoppers get rated poorly for this action which in turn significantly damages their earnings, and Customers are left frustrated with their Instacart experience.
Frustrated Customers means brand loyalty decreases which incentivizes Customers to leave Instacart and consider competitors to meet their grocery needs.
To figure out why the utilization rate of the refund/replacement system was low, a better understanding of Instacart Customers’ relationship with replacements and refunds was needed. Relevant experiences that were shared online by customers across Youtube, Twitter, and Reddit were gathered and synthesized with an affinity diagram alongside six interviews with Instacart Customers.
.png)
While the initial organization of the data confirmed and highlighted the various frustrations of Customers, a closer look revealed that Customers actually had Shopper preferences. This meant that when they were matched with a Shopper who aligned with their preferences, Customers were happy with their Instacart experience. Conversely, a mismatch meant a negative Instacart experience.
The discovery that Customers had Shopper preferences led to the revelation that Customers’ relationship with the replacement/refund system was actually intentional. Customers knew all along that they could provide instructions for Shoppers but they purposefully chose not to because they preferred when Shoppers communicated with them. Other customers were diligent users of the refund/replacement system and got upset when Shoppers communicated with them or strayed from their instructions.
Restructuring our affinity diagram to reflect this divide clearly exhibited that there were two indisputably separate user groups:
.png){kind=small}
.png){kind=small}
The existence of these two different user groups was entirely responsible for the low utilization rate of the refund/replacement system by Customers, the frustrations by Customers with Instacart, and the subsequent low ratings and pay Shoppers receive.
This meant that the original problem of low utilization by Customers of the refund/replacement system hurting Shoppers and Instacart was in fact merely a symptom of a larger problem which is that Instacart was designed to be of service to a singular type of user – one who would use Instacart perfectly. This was the real problem.
Since a closer look at customers’ interactions with the Instacart replacement/refund system revealed two separate user groups for Instacart, we decided to see if other players in the online grocery industry recognized and successfully accommodated these two distinct groups in their own replacement/refund systems.
A competitive analysis illustrated the competitive differentiators of major online grocery delivery services and if Instacart was losing a significant market share by failing to accommodate for these user groups.
Research indicated that all competitors either supported these groups with a singular replacement control or had a different service altogether, which confirmed that there was a clear opportunity for Instacart to grow their market share.
Being a cross-platform service, Instacart can be accessed via website and mobile app. Since two-thirds of online grocery orders are made from mobile phones⁴, we opted for a mobile solution for ease of subsequent testing with Instacart Customers.
Additionally, Instacart has both a Customer and Shopper interface – each their own separate app. Given that Instacart Customers and their interface were more readily accessible to us than Instacart Shoppers and their respective interface, we decided to focus our solution on the Customer interface.
Ease of testing with Customers due to two-thirds of online grocery orders being placed on mobile
Customers and their interface are more readily accessible than Shoppers and their respective interface
Since the heart of the problem lay in the existence of two user groups, our solution had to appropriately recognize and represent the needs of these unique users. With this in mind, we sketched out a solution with the following:
All customers can conveniently specify refunds/replacements/item notes in one spot from the product page without navigating to the cart and tapping multiple times to achieve the same effect.
.png){kind=small}
Unassisted Customers can now quickly specify "No replacements" for all items with a single tap instead of selecting for each and every item. Shoppers are also explicitly notified to issue only refunds for out of stock items.
Allows all Customers to communicate to Shoppers which shopping experience they prefer, which then informs Shoppers on how to best serve each Customer, leading to satisfied Customers and decreases in low Shopper ratings.
To prioritize measuring for the solution’s effectiveness, it was agreed that a high fidelity prototype – despite being more costly in terms of time and effort to produce – would most accurately capture the behaviors of real Instacart customers. We evaluated three different prototyping methods.
A “one size fits all” prototype designed to be interactive for a specific flow. Participants would follow a specific set of shopping instructions and asked about their experience and thoughts afterwards.
– Cheaper to produce
– Forcing participants to adopt foreign behavior
– Can't trust what users say, only what they do
A static prototype where the participants' interface is controlled by the designer. Participants' clicks are inactive so the designer displays the screen that correlates with their "click."
– Users are more free to shop "naturally"
– Generates more reliable and realistic data
– More expensive to produce
– Not as realistic as letting participants navigate the design with their own devices due to breaks and stutters in the flow
– Risk of human error
An interactive prototype with limited functionality. Participants would be instructed to purchase specific items, but are otherwise free to shop for these items as they wish. Because this prototype has limited functionality, participants would be notified when an action is not possible and asked to verbalize their actions.
– Opportunity for participants to make choices and “shop” naturally
– Generates more reliable and realistic data
– Not the truest representation of shopping behavior
Of the three options, the Limited Functionality prototype appeared to be the best compromise of the three. Given the fact that all three options weren’t fully capable of capturing realistic customer behavior, the Limited Functionality prototype represented the most favorable tradeoff between the resulting data versus the time and effort required. While it would take more time and effort to create than the One Size Fits All approach, the Limited Functionality would be cheaper to produce than the Wizard of Oz model. With this decision, a high fidelity prototype with limited functionality was created using Instacart's native design system.
Remote moderated usability sessions were conducted with five customers to observe utilization rates of the new replacement/refund system, "Shopping Preference" feature, and overall order satisfaction.
* This is to due to the fact that one participant always preferred replacements for their order so this feature was not relevant to them.
Many of Instacart's criticisms appeared straightforward, such as the low utilization rate of the refund/replacement system by Customers causing the low ratings and pay Shoppers receive. But as our research revealed, the neglect of a fundamental user group lay at the heart of these issues. Our additions of a more prominent and consolidated refund/replacement screen, "No replacements" checkbox, and "Shopping Preference" feature account for this omitted group and reconcile the criticisms from users and the media while discovering a competitive opportunity.
Bringing the refund/replacement system earlier in the checkout process serves as a visual prompt for customers to provide instructions, which helps Shoppers better shop their order, which means higher order satisfaction rates. Plus, having all your options in one place makes for easier decision making.
.gif)
An easily accessible global "No replacements" checkbox supports the needs of the omitted user group by allowing for a quick checkout and ensuring no unwanted items when they receive their order.
.gif)
Enabling customers to select their "Shopping Preference" ensures customer order satisfaction rates, decreases support costs, and secures Shoppers' livelihoods.
.gif)
Implementing these validated features into Instacart’s product not only lends a secure foothold in a burgeoning industry, but also lends the following advantages:
A more prominent and consolidated refund/replacement screen allows Customers to more easily add instructions which would lead to Shoppers purchasing replacement items instead of just refunding when customers aren’t available for communication. Increased purchases of replacements would naturally contribute to increased revenue for Instacart.
The addition of the “Shopping Preference” feature would be entirely unique to Instacart. Its novelty and practicality immediately communicates value to online grocery customers and establishes a compelling competitive advantage for Instacart, which opens the door for competitors’ customers to switch to Instacart.
The global “No Replacement” checkbox allows Instacart to offer the same support that their competitors have already been offering for a while for Customers who want an unassisted shopping experience, which can make way for an increased and stronger market share.
Apart from the increases, the client can expect to see these reductions:
Fostering happier customers can lead to lower order refund costs. Lower refund rates reduce the amount of support staff needed to handle complaints, which also lowers support costs. Lastly, happier customers means there's a lower likelihood of losing customers.
Our proposed solution makes for a more inclusive and thoughtful Instacart experience for everyone and prepares Instacart to face the market straight on as we head into a booming era of online grocery shopping.
1. “Dec: U.S. Online Grocery Sales Total Nearly $98 Billion for 2021.” n.d. Www.brickmeetsclick.com. Accessed February 11, 2022. https://www.brickmeetsclick.com/dec--u-s--online-grocery-sales-total-nearly--100-billion-for-2021.
2. “US Online Grocery Sales Growth Set to Reaccelerate as COVID-19 Concerns Grow.” n.d. Www.spglobal.com. https://www.spglobal.com/marketintelligence/en/news-insights/latest-news-headlines/us-online-grocery-sales-growth-set-to-reaccelerate-as-covid-19-concerns-grow-65856919.
3. “‘Instacart, We Estimate, Controls about 25% of Online Grocery’ Market, Says Baird’s Colin Sebastian.” 2022. CNBC. March 7, 2022. https://www.cnbc.com/video/2022/03/07/instacart-we-estimate-controls-about-25-percent-of-online-grocery-market-says-bairds-colin-sebastian.html#:~:text=TechCheck-.
4. “Grocery Shopping Goes Mobile” n.d. Www.retailwire.com. https://retailwire.com/discussion/grocery-shopping-goes-mobile/
