Written by Author Name 1/30/16 | 4 Comments

Data-Driven Marketing Strategy

In the multi-channel, cross-device world it can be a challenge to measure the real value of a marketing channel.

To make sure that you correctly understand which channels are your defeaters and which are your attackers because if you don't do so, you're gonna get some channels which will never bring it to the final goal of your performance marketing which is conversion. 


To understand this conversion here’s an example: when travelling from Berlin to Wroclaw you obviously have a wide choice which mode of transportation to take -  flight, train, carpooling or bus. To make a correct ticket purchase you go through some aggregator websites, go directly to brand websites of big carriers. In the end, you buy a ticket but if you counted backwards, there would have been five websites or so, including both branded websites and aggregators which you visited before. That is an example of the multi-touch user story because, in order to pTo ensure your marketing team is effectively contributing to your business, you need an actionable, data-driven strategy. Data can allow you to make strategic decisions helping your marketing efforts become more successful.

At Touch Digital Summit 2019 our speaker Anastasia Stefanka, a Business intelligence analyst from BlaBlaCar, talked about the marketing strategy at BlablaCar, how they are using data for it and to what main pillars they are concentrating when trying to analyze their marketing mix.


BlaBlaCar is one of the French unicorns (a private company with 1bln+ valuation), an online marketplace for carpooling. It’s the world’s leading long-distance carpooling platform which connects people looking to travel long distances with drivers heading the same way, so they can travel together and share the cost. Their vision is to bring freedom, fairness and fraternity to the world of travel. 


In her talk, Anastasia went through common pitfalls in data-driven marketing and discussed main approaches on how to identify value-adding channels based on the goals of marketing, and how to turn data into insights. Prior basic performance marketing knowledge is recommended, but not mandatory.


BlaBlaCar as a marketplace is focusing on promoting its products and this is why today we are going to be talking about performance marketing and data-driven strategy in performance marketing in three main pillars.


ATTRIBUTION 

Probably for everybody who is doing marketing (performance-oriented marketing in particular), the question of the return of investment is something which is on top of the mind of all the marketers. We are trying to optimize our marketing channels and marketing mix on a daily basis.


As Simo Ahava once said: “digital marketing is teamwork, where points are awarded for blasting through competition in an attempt to reach the goals we set for our clients”. 


What stands behind this quote actually is that if you want to have an efficient marketing strategy, you look at your channels as team players that have to score a goal and each of the team players contributes differently to that goal. Naturally, you have somebody in your team who tries to protect your own gates, or somebody who is good both at attack and protection, or somebody who is a very good attacker actually trying to score that goal. 


So the idea of the marketing mix is tourchase a single ticket, you had to go through numerous websites to make your comparison, to make a choice and by the end of the day you have been influenced by multiple marketing channels and multiple advertisements as a user. And actually it happened even on multiple devices. 


So the question is from these five websites which one, in the end, made you decide and this is not a trivial question to answer. It's basically the point which all the marketers are trying to answer as we speak because when we talk about attribution, we want to assign a conversion (the purchase of a ticket in your case) to a channel which was effectively the most useful for you in purchasing. But how to evaluate this? There are multiple strategies. The most common one is the last click, so the last batch point wins, the last website which you visited was the one which convinced you, but there are also multiple points to consider there.


The topic is quite complex for most of you who are selling some digital services online, be it conversions, branding stories, etc. What matters in the end that this branding story is a conversion and to measure the conversion correctly you have to apply a rule which you're going to make for all of your marketing efforts.


When using a very broad marketing mix both online and offline, we look daily at different sources which we try to optimize based on their performance in terms of return on investment, how many conversions do they bring us, how much do they cost us, is it worth continuing this effort. So the point which you should take is to apply a rule-based model, for example, which says you're going to attribute your conversion to the source which participates in your user purchase funnel and try to look at these sources and compare them when it comes to cross-device. 


So if you're considering developing your own attribution model which factors in cross-device, this could be something important for you as well. Think of your marketing channels as a team which all try to get your user to the point of conversion and then once they reach it, you know which player was the most efficient. This is the basic rule-based model. It’s something you can start working with already today in the sense of your own conversions because if you don't do so, all the big players on the market will do it on your behalf. Big marketing channels like Google, Facebook is actually quite greedy and they all try to assign conversions to themselves. So if you're gonna just simply sum up the conversions in your Facebook and Google account, well you are going to be disappointed because there is going to be more in that sum than an actual amount of purchases from your website. That's why unless you think of it yourself, they're gonna think of it for you and they're gonna try to convince you that they are the most important. That's also our first point - attribution which means thinking of your channels as a team that plays to score a goal and assigning a rule to your team under which a particular player will be the actual goal scorer.

MEASUREMENT

Let's imagine you are in a situation when you have developed our attribution model and finally see that there are certain marketing channels that are bringing you more conversions and other marketing channels bringing less. Well, the first thought you would have, would probably be to invest more in something that delivers more conversions and less in something that delivers fewer conversions. That could be true to an extent but let's think of it from the perspective of the purchase funnel of the user.

The measurement point is about the purchase funnel of the user. Users can go through multiple stages starting from awareness when you need to convince the user that there is your company out there on a market, then there comes consideration, intent and finally purchase

Imagine that you are playing with this game and you are doing the conversion scoring game. At first, there comes some kind of YouTube campaign and there the user will start knowing about your company and after that, he will click on the search advertisement and is going to convert. Has YouTube participated in the conversions? Yes, it did. Was it the last one? No, it wasn't. So maybe Google searches what you need, but would the user actually have searched for you if he hasn’t seen a YouTube ad in the first place. That's a good question and that's why comparing marketing channels just by the total number of conversions they each score is not exactly perfect. 

You all know the famous meme about our education system when each animal (be it an elephant, monkey, fish, penguin, etc) has to take the same test to climb a tree for fair selection. In terms of marketing channels, it's ideally not the case that you are comparing your marketing channels only by the number of conversions that they are bringing. While doing so, you're undermining the channels which are staying upper in the purchase funnel and you are over-evaluating the channels which sustain lower in the purchase funnel, for example in such case YouTube is a fish and search is an elephant. You would always assign conversion to a monkey, for example, so research would be a monkey in that case. You would always assign conversion to a monkey, but never to a fish which was also quite helpful in the beginning.

All the big guys over there are trying to invent a way to measure the effect of the upper funnel advertisement which it has on the lower funnel advertisement. You can use the incrementality methodology to predict how many more conversions the user gets from all the channels if they're exposed to some upper funnel channel. In our case would the user have searched for a ticket from Berlin to Wroclaw if he didn't see a YouTube advertisement in that case. So that would be the incremental value of a channel and if you talk measurement, then thinking of your team of the players is not only an absolute amount of goals which they are scoring, but it's also the number of helpful passes which they give to other players to score the goal. This is the idea of measurement and incrementality.

LIFETIME VALUE

As mentioned earlier this purchase funnels of awareness, consideration, intent and purchase. So what happens here. You made your team. On each stage you have a player, you have a marketing channel which is covering each of the stages of the purchase funnel, they all help each other score the final goal. It looks pretty much like this and then the goal is scored - the user purchased, the conversion happened. 

So afterwards you want to maximize conversions. What would you do? Would you go through all the purchase funnel for the new user again? Well, that makes sense. That is the purpose of all the acquisition activities. That's why they're called acquisition because you are acquiring new users through the funnel. But what you can actually do is think about that it would be cheaper if you could take the same user who already went once through your purchase funnel and just kick him back into another purchase funnel. But as he already knows about you, this means that you can start at a lower stage and spend less money on it. For the competitive markets, you can start pushing the user to consider your product again this sense of retention and for the markets with the lower competition, you can even try to generate more intent by cross-selling. This is what the retention is and this is how you can try to reacquire users, basically trying to make them come back to you again. It can be a challenging thing especially depending on the type of your product. That's why cross-selling exists. 

The logic here is that it's always cheaper to reacquire already existing users because they already went through all of the steps with you than acquiring a new user who has no idea what your product is about and who needs to be convinced from scratch. Convincing from scratch is expensive and takes time. Reconvincing already a user who has already been at the bottom of the funnel is easy and this is what you can do. 

From here comes the concept of lifetime value which in its nature is how much of the user brings you over its lifetime. Lifetime is a very stretchy term, so you can apply some time limit onto it like one year or six months, you decide to depend on how long you are willing to wait. So in a nutshell, it shows how much you gain through the user with repetitive purchases and with that, you have a concept of cohorts, for example, when you can look at the different users acquired through different channels and see if their lifetime value is same or different, do the users whom you first acquired through Facebook come back as often as users whom you acquired from Linkedin or maybe there is a certain pattern which you can notice in terms of a source of acquisition and cohorts. On the other side, it could also be the sense of seasonality, for instance, you can observe seasonality and users acquired in lower seasons (where there are fewer people active) would be more valuable to you than the users acquired during the high season.

Watch Anastasia's full talk here