Use-case optimisation lowers platform costs by 15 %

JustPark is a park and pay app – a sharing economy digital platform akin to what Airbnb is in the real estate world. It links 5.5 million active drivers with over 45,000 listed parking locations in the United Kingdom via digital touchpoints.

JustPark offers an opportunity to individual space owners that have an extra parking space or two. They can list these online with JustPark to establish an additional stream of income. It also offers a great additional revenue stream for big parking space owners such as hotels and churches that have plenty of parking space with variable demand. It offers significant improvements for the end users too. A driver can book a parking space in advance and in minutes without the need to use cash in the comfort of their car and their smartphone.

To handle the personalised and automated communication across its various channels with a large and active user base, JustPark relies on Bloomreach Engagement, a CDP and Automated Marketing platform. This platform unifies behavioral data from the app, the web, marketing communication, driver’s vehicle data and bookings into a single customer view. In addition to the data tracking aspect, Bloomreach powers many automated email and mobile notification flows as well as JustPark’s campaign reporting, A/B testing and in-app targeted “pop-ups”. This tech-stack is new and there aren’t that many specialists available to hire and JustPark felt that it wasn’t using the tech stack to their full potential.

Colours of Data and Datacop were brought in to JustPark to audit their use of their CDP tech stack with the aim to see how to maximise the value of the investment into unification of customer data. This involved the evaluation of the existing data architecture and use-cases for which Bloomreach was being used and then the identification of issues and our recommended solutions. The audit showed that the improvement of one of the existing use-cases in particular - the so called “4-Week Vehicle MOT Expiry Reminder” flow could bring the highest amount of value in the shortest period of time.

This use-case is a flow sends 4 weeks, 2 weeks and 1 week before a user’s vehicle’s MOT certificate expires an email as well as a mobile push notification. If the user clicks through on either the email or the push notification - they are redirected to a personalised landing page that has pre-filtered results based on the user’s vehicle and home location information provided to JustPark by the user. This saves time for the user as well as reminds the less-organised car owners to about their MOT responsibility.

Based on our assessment and in close cooperation with JustPark’s Marketing and Growth Teams we have managed to deliver 3 significant improvements to the use-case:

1. Re-designed the vehicle data pipeline to ensure 100 % eligible users receive the campaign.

Problem/Situation: The use-case’s original design did not take into account that it will be running for more than a year. The logic of the data pipeline that fed the CDP with vehicle information -
such as the time of the year when that particular’s vehicle MOT certificate expires, would not import the renewed MOT expiry timestamp in case the user’s vehicle has been registered with JustPark with more than a year.

Data Analysis: We have confirmed this with JustPark by comparing expected send outs as based off data from the source database that collected the external vehicle information via a data partner and the actual send outs in the tool. After a year of the use-case running, the number of send outs started to drop dramatically.

Implementation: We fixed the API Webhook to be capable of rewriting the stored mot_expiry_timestamp if the vehicle exists in the database and made sure the SQL import is using the time attribute which refers to the latest updated mot expiry timestamp and not based off its vehicle creation timestamp. Approx. half of users who should have been receiving this were not and when the fixes were implemented, they returned to their normal level.

2. Re-designed the automated flow logic to use 60 % less data capacity with the same volumes of send-outs.

Problem/Situation: The use-case’s original design was overly reliant on API Webhooks to transform data into their needed format for the automation to function. That created 2 extra events during the automation process that are not necessary and can be replaced with native features of Bloomreach Engagement.

Data Analysis: Identifying all the events that are necessary and the ones that are not necessary throughout the duration of the flow was active, that would have been approx. 60 % of all the events that do not need to be tracked.

Implementation: This required the changing of the flow logic and the setting up of customer data attributes using Bloomreach Expressions and Scenario Filters. Because CDP’s costs are directly linked to the number of events processed by the system, this had a major impact on the annual cost of the software for JustPark. Stopping the tracking of these events represents up to 15 % of annual platform costs for JustPark.

3. Growth-hacked the email creatives to increase their CTR % of the campaigns by +15 % and +11 %

Problem/Situation: The use-case’s original visual and copy proved successful in proving the value of the concept. Since the original visual and copy design was minimalistic, we did a creative brainstorm with JustPark’s Growth Team and made a list of potential interventions. From this list 2 ideas had the highest perceived impact and were reasonably quick to implement given the resources available.

Data Analysis: To evaluate the impact of the proposed interventions we have conducted 2 major A/B tests. The first was testing the viability of an attractive and thematic hero image. The test involved 3 variants. Variant A contained a Hero image focused on a car. Variant B contained a Hero image focused on a garage.

After the first test concluded, we designed a second test that involved 2 variants. The tests were split because the size of the sample size wasn’t big enough to run 2 concurrent interventions at once and receive statistically significant results in a reasonable time. Variant A contained a copy that invoked the idea of “Social Proof” - people are more likely follow your nudge if they are made aware that others feel similarly (or act similarly). The other variant was the control group with the original copy.

Results: The car visuals outperformed the control group by +8 % CTR and the garage visuals had an uplift of +15 % CTR compared to the control group in a 3-week testing period. This has shown that having image visuals in this automated scenario has value and set all the emails with the Garage visual. We agreed that going forward a test using an animated Hero image vs a static Hero image could be an interesting to drive incremental growth further.

Emails with the Social Proof Copy had a +11 % CTR uplift in a 2-week testing period. This has shown that using a well phrased Social Proof in an appropriate context has a positive impact on CTR of an email.

2. 8. 2022

Our projects