Author: asakala

Protected:

This content is password protected. To view it please enter your password below:

Tesco

Tesco
Shopping App & web design team

1. Personalised recommendations
2. Post order and day of delivery
3. Voice of the customer
4. Smart categories

1. Personalised recommendations

How might we facilitate an easier base shopping experience so that we allow customers the time and space to discover items they might not usually buy?

Team

Recommendations API – Sahil
Data Science – Stuart / Guerrmo
Front end engineering – Jon / Ruthwik / Priya
Mango API engineering – Ateen
Product Owner – Megan / Emily
Product design  – Asakala

Metrics to watch: Combined order conversion, items per order, average order value

Impact

Show me relevant products based on what I have just added to basket

By showing customers relevant recommended products based on what they have just added to basket, basket size will increase as friction to shop will be decreased allowing time and space for shopping items not usually bought.ow me relevant products based on what I have just added to basket

Assumption – customers have a finite amount of time to shop. By reducing friction on the “base shop”, this will free up energy for more inspired shopping.

Exploration around how “based on this choice” suggested products could appear – based on assumption that once customers have added to basket, the rest of the products on the listing page are no longer relevant.

If this appears on every add to basket how can we mitigate any annoyance this could also cause to customers
Have it appear x # times – and if the user hasn’t clicked on it – don’t show again?

Ideas for how the interaction could work

Show me relevant groups of products based on what I have just added to basket
How might we encourage customers to go beyond their normal shopping routines and inspire them to try new meals and products based on items they already buy?

Where in the journey might we surface personalised recipes?

//player.vimeo.com/video/441066198?title=0&byline=0

Refined Ideas to A/B Test

Returning Shopper
Home page personalised products carousel > customers adds to basket > based on choice recommended product component appears

Hypothesis – by showing customers relevant products on the home page, friction will be reduced as customers will be able to browse relevant products more easily.
By encouraging customers to shop from recommended items based on what they just added to basket using the “Product to Product” algorithm, customers will shop faster and more easily whilst also driving profit from higher Margin items.


Search PLP
customer searches for product > adds to basket from search results PLP > based on choice recommended product component appears


Hypothesis – by encouraging customers to shop from recommended items based on what they just added to basket using the “Product to Product” algorithm, customers will shop more easily as friction will be reduced whilst also driving profit from higher Margin items.

Caveat: this will only be active for products which are “exclusive” in that customers only add 1 items from search results PLP.

Metrics to watch:
Adds to basket from recommendations versus search


1. Personalised recommendations
2. Post order and day of delivery
3. Voice of the customer
4. Smart categories

2. Post order and day of delivery

How might we help customers feel informed and valued on day of delivery?

Team

Van tracking API  – Mahesh
Data analytics – George
Front end engineering team
Mango API engineering team
Product Owner – Louis
Product design – Asakala

We believe that reducing anxiety after a customer has placed their order will:
– reduce calls to customer service
– improve customer satisfaction
– increase customer retention from their 1st to 3rd shop.

The moment when an order arrives should be the moment of delight in this journey to cooking and eating. We know from call data that customers are anxious about the delivery whether it is late or on time. Can we use research around the psychology of waiting to mitigate frustration and help customers feel supported.

Hypothesis

Showing customers an eta is only relevant and accurate when the can is the next stop or when we estimate we will be late – and is more in line with current tracking technical capabilities.
A better day of delivery experience will improve loyalty from 1 -3rd shop.
Metrics for success: CTO (calls to our customer engagement centres) and star rating feedback.



1. Personalised recommendations
2. Post order and day of delivery
3. Voice of the customer
4. Smart categories

3. Voice of the customer

How might we better utilise the feedback given by customers?

Team

UX Research – Daniele
Product – Jack
Product design – Asakala
Engineering – OpinionLab

Impact

By a simple prompt, customers changed their mind and more than 8 million customers chose to change their mind and not use plastic bags! 🐳
Asking customers during checkout didn’t have any affect on conversion.
Insights into how the copy on a CTA can influence behaviour.
A wealth of qualitative data was collected around customer’s habits and opinions.

Validate assumption that allowing customers to give feedback will not affect order conversion – let’s use another problem space in the business as a test scenario!

During checkout customer who has carrier bags preselected or chooses carrier bags > additional modal appears with text and CTA for feedback > feedback pop up appears

During checkout customer who has carrier bags preselected or choos

Hypothesis
On the order summary page, if we ask for customer feedback to those customers who usually have the “pack with bags” option selected/choose to select “pack with bag” – then at least 5% of customers will leave feedback with no negative impact on Order Conversion rate.

The most successful variant

While all the variants managed to increase the number of customers that switched to packing “without carrier bags”, Variant C’s combination of a “Tell us why” call to action button, along with the explanatory “plastic footprint” text clearly performed best in both generating the initial survey interaction, as well as in form completion.

These results show that customers can be persuaded to switch to Bagless even when in the middle of their shopping mission, without any negative impact on order conversion rate.

We recommend the “tell us why” CTA and simple explanatory text be adopted as the baseline wherever possible when soliciting customer feedback in the future.

Given the present day focus on climate change and the negative impact of plastic waste on environment, we may see an even bigger switch from bagged to bagless in the near future.

1. quick jump to__Personalised recommendations
2. quick jump to__Post order and day of delivery
3. quick jump to__Voice of the customer
4. Smart categories


4. Smart categories

We have a large increase of new users on App, however conversion is much lower than expected and compared with web.
Assumption based on observations during user testing: shopping on App takes longer and so customers find it harder to build a basket.

How might we facilitate faster product discoverability on App?

Team
Android engineering – Darmesh / Swathi / Sudanshu
iOS engineering – Bittu / Jagadeesh / Adarsh
Product – Megan / Emily
Data Analytics – Bruce
Product design –  Asakala

Metrics measured
time taken to shop/ average order value / page views per add / items per order / browse conversion rate / revenue per visit

Impact
Whilst we saw no affect to the initial order conversion rate we saw a very small lift in IPO (items per order) and a reduction in the page views per add which suggests that customers find it easier to add the products they want to buy – this was for the Variant B.

Customers currently need to click through 4 layers to get to bananas – our most sold item!

How exactly does the current taxonomy currently work?

Design development

1. Is the prototype code ready and ready to implement into A/B test? Would this just be on Android initially?
2. What algorithms are we using for this initially?
3. Can we hard code in the imagery to “fake” the test without having to connect the back end system?

Smart categories – How might we surface products earlier to customers so that they might browse the taxonomy faster?

Helping customers to build their baskets faster
By surfacing 2nd and 3rd layer taxonomy categories to customers earlier, conversion will increase as customers will be able to browse relevant products faster. We know that out of customers who don’t convert, 57% have baskets below the £40 minimum basket value

Opportunity to tailor how we sort our products
By using the personalisation algorithm combined with other logic such as sorting products by margin, we may increase profit by using more integrated ways of showing the existing recommenders to customers.

Helping First Time Shoppers
By surfacing popular categories (customers add the most to basket from) we can help first time shoppers who don’t have favourites to build a base basket faster.

Metrics to measure
time taken to shop/ average order value / page views per add / items per order / browse conversion rate / revenue per visit

Navigate

1. quick jump t0__Personalised recommendations
2. quick jump to__Post order and day of delivery
3. quick jump to__Voice of the customer
4. quick jump to__Smart categories

Back to the top        Home

End of tape  – thank you for your time
Any questions feel free to contact me on

☎️+44 (0) 7725 988612
✉️info@asakala.com / asakalageraghty@gmail.com

View my current CV