What difficulties did we encounter?
These are difficult times for importers, dealers and buyers. Demand for cars is declining. This is facilitated by both geopolitical factors and a decline in consumer power.
The situation is aggravated by the demographic pit of the 1990s-2000s. Our target audience is people aged 25 to 45. But now this audience is shrinking. And therefore the price of traffic is growing.
So we realized it was time to change our approach and tighten the “fight” for each client to get the same and even better results on a shrinking audience. We need to change our approach albania phone number resource pay closer attention to marketing and — especially — analytics.
How we work with data
There are several traffic sources that we look at — Google Analytics, Yandex Metrica, and others. They are accumulated in ClickHouse. CRM data is also pulled there, from offline. We pull traffic costs from advertising campaigns — online accounts.
All of this goes through the open source PostgreSQL database management system. It allows you to create and store databases, and work with data through SQL queries. It also includes information from the Calltouch analytics service about interactions along various user paths. This information is transformed into web data marts and final reporting.
What problem arose?
At the moment, analytics in the automotive industry is not the most perfect. How do we see the user? He follows the target request, does not interact with the site in any way. Therefore, the user disappears from our field of vision until the contract appears offline. But with such analytics, managing marketing is like driving a Zhiguli at night without brakes and lights.
In search of a solution, we paid attention to related industries. For example, in e-commerce, the story with the ROPO indicator is very strong. Research Online Shopping Offline - which translates as "study online, buy offline". The ROPO effect is a phenomenon when people find information and compare prices on the Internet before making a purchase in a physical store.
Colleagues have been working with the flow of traffic from online to offline for a long time and have learned to analyze it. There are many studies that confirm that about 80-90% of users first study a product online and complete the consumer cycle offline.
How to evaluate the ROPO effect
Let's talk about solutions that allow you to evaluate ROPO. First, I'll list the methods that don't work very well:
Surveys. The main disadvantage of this method is that when asked "how did you hear about us?" people usually answer "from advertising" without naming the exact source of traffic.
Sales volume. This is when we launch an advertising campaign and see how it affects sales. The accuracy here is also low.
Wi-Fi traps in dealerships. They can be installed in different places to see how users intersected - this is good. The downside of the method is that not everyone turns on Wi-Fi. The accuracy is lame.
Fiscal data operator (FDO). For FDO analytics, it is necessary that payment be made by bank card, but most purchases are made on credit. It does not suit us.
Mobile operator data. There are many operators on the market, each with its own piece of the audience. Information about users will have to be purchased from different sources.
Visitor counters. The most inaccurate method: the system will even record managers.
We have chosen the most relevant tool for us – QR codes.
For example, at a dealership, each used car has a QR code with a price tag and additional information. Let's be honest, not all customers scan them - many immediately call the seller. But those visitors to the dealership who use the code get to our site - to the car card with a tag. Thanks to this tag, we can unravel the tangle of all the user's visits and find out from what sources he came to us.
As a result, we saw that about 30% of those scanning QR codes had already been to our sites, but had not interacted with us online.
How We Studied the Customer Journey of Offline Users
The Big Data tool from Calltouch helped with this , which allows you to track the history of visits to web resources by phone number. We collected the contacts of 42,246 buyers who did not show up online and ran them through the Calltouch service . The results are below.
30,885 people came to us through classified ads, 2,150 through Yandex and Google maps, 9,211 did not leave any digital trace, that is, they came offline.
We also looked at how these users interacted with our paid online advertising. Data was collected in two areas - used cars and new Hyundai cars. It turned out that more than 20% of customers - no matter whether they wanted to buy a new car or not - visited the site rolf-probeg.ru.
I would like to focus on the place that sites occupy in the touch chain. The bulk of traffic comes from the first touch. At the same time, there is a long tail of touches where users and audiences of different sites intersect with each other.
Using the graph below, we have shown how exactly and between which sites they intersect. Despite the high volume of landings, promo pages, location sites, sooner or later people still come to the main sites of the company.
How many touches and time does it take to make a purchase
We answered the question of how many people come from offline, but continued the research. We were interested in how users interact with our resources and after what touches they perform target actions - this is an application, a call, creating a worksheet.
The bulk of target actions occur during the first contact with the site. At the same time, more than 50% of buyers interact with our resources more than five times. Next, we measured the time from the first contact to the decision.
It turned out that 1-5 days are enough for only 8% of users. This suggests that the audience is not ready to buy a car now, since this is a fairly large and expensive purchase that requires more detailed study. 97% of visitors to our dealerships need 91-100 days to make a purchase decision.
