We have been talking about one-to-one marketing since the 90s. But the companies that have been able to really implement it are few and far between. Insurance company InShared has made significant strides in recent years. Marketing manager Bettina Dalenoord talked about it during the ‘Conversion Power Session’, organized by Webpower, Magneds and ConversionMob.
InShared is a new player on the insurance market that has managed to attract many clients due to the fact that they give back money they don’t have to spend on claims to their clients. Television commercials have greatly increased brand awareness, but the real transaction is taking place online. In order to better capitalize on the individual customer journey, InShared joined forces with Webpower partner ConversionMob four years ago, which is a digital marketing agency focusing on maximizing the returns on every digital interaction with a consumer.
When InShared started the process, their marketing was actually very traditional. The organization underutilized the power of online marketing; the fact that everything can be measured and analyzed. Marketing manager Bettina Dalenoord explains: “We’re always thinking in terms of the same formula: traffic x conversion x margin = total margin. When you introduce a new brand to the market, you first focus on generating traffic. But this traffic won’t do you any good in the end if your conversion is lagging behind.”
With her goal of increasing conversion in mind, Dalenoord went to ConversionMob with four questions:
- How can we improve conversion attribution, allowing us to better determine which channels to use our marketing budget for?
- How can we get a grip on ‘cross device behaviour’?
- Which prospects who end up on our site are most likely to convert?
- Which channel should we use for which prospect next time in order to increase the likelihood of conversion?
To start with, ConversionMob and InShared carried out a data audit: which data are we currently saving, and what information would we like to have? Can we use the data with our current method of saving data?
After this foundation had been set up, the team could start focusing on the first question: how can we improve conversion attribution? InShared used a conversion attribution model focusing entirely on the last-used channel (the so-called last click model). “Of course our instincts told us that prior customer contact also contributed to conversion. The question was: how can we assign the right value to all channels the customer uses during their customer journey?” says Dalenoord.
She uses a fictional customer journey as an example. The initial contact with the customer took place on social media. This led to the customer signing up for the email newsletter. They used Google when they started looking for a specific insurance company, and ended up at InShared via SEA. They did not convert during that first visit. That happened a few days later when he visited the website again, this time through SEO.
Within the world of attribution, there are seven standard attribution models. Which of these models you use depends on the behaviour of the ‘average customer’. If the average path to conversion is relatively short, the differences between some of these models are minor. By adding additional elements, like the engagement of a visit, or the time in between visits, you can optimize a standard attribution model. If it turns out that contacts established through the newsletter last longer, with people viewing more pages on your website, you can assign greater value to this channel.
Dalenoord: “Eventually, each channel’s contribution for each customer will end up being different. You simply need to choose a model that most closely approximates reality based on the information you already have regarding customer behaviour. And that’s something you can only achieve through testing and optimization. In the end, your goal is to gain greater insight into which channels contribute the most to conversion. And this may be different for every campaign.”
Building customer profiles
The next step in the process was predictive profiling. InShared works with three main groups: quick deciders, average deciders and people who do a lot of research before they take out insurance (the heavy searchers). The behaviour of these groups varies greatly. A quick decider, on average, visits less than two websites in order to make a choice, while a heavy searcher will visit almost 30. It goes without saying that you need to approach these two profiles in completely different way. The quick decider wants to see an action page immediately, without having to wade through a heap of product information, whereas a heavy searcher will be intimidated if the step to conversion happens too quickly.
Dalenoord: “Because we can assign prospects to one of these three profiles really quickly based on their clicking behaviour, we can more accurately interpret their behaviour and act accordingly. In addition, we are able to predict which people will end up becoming customers within a 90 percent confidence interval. This enables us to better focus our marketing efforts on the prospects with the highest probability of converting.”
InShared learned something else as well: “Our website was designed in such a way the it appealed mostly to the heavy searchers. We provided a lot of product information, but the short route to make a quick decision had not been optimized. We are currently researching how we can adjust the website in such a way that we can appeal properly to all groups. And we’re looking at how we can change our advertising and email marketing campaigns to better match the customer type.”
In the future, the team also wants to get started with linking attribution and persuasive profiling, as well as tackling cross-device behaviour, but they’re careful about moving too fast. “We’re using a step-by-step approach, based on agile principles,” says Dalenoord.
InShared gained many insights over the course of the process. Dalenoord names a few. “Back when we still did conversion attribution based on the last-used channel, we thought a large part of our SEA investments had a negative ROI. This is why we stopped SEA at the time. But we are once again using this channel. The situation is completely different now, because we’re using a different conversion attribution model. The same principle applies to determining what affiliates we work with. In the past, we always thought the affiliates that generated the most sales were the best ones. But this isn’t necessarily the case. Some affiliates are really good at conversion, while others are good at providing new prospects. And this is incredibly valuable as well, because once those prospects are in our database, we can start communication from there, making sure they convert. This means we need to tweak the affiliate attribution model as well.”
Her opinion on price comparison sites has, likewise, changed. “They were our most important channel for car insurance policies. But now that we know much more about customer behaviour, we’ve notices that comparison site mostly direct people to us who know us already. Our goal now is to make sure those people visit us directly, instead of through a comparison site. We’re investing more into our own channels rather than external ones.”
She already mentioned the most exciting thing: “We know with 90 percent certainty whether someone who visits our website is going to take out an insurance policy or not. Just imagine what that means! We still need to discover in what ways we can actually use this knowledge, but it’s clear that we know so much more about who to target with our marketing efforts.”
Naturally, Dalenoord can hardly wait to take the next steps in this process. “It’s truly inspiring and motivating, almost addictive. Each new insight leads to a follow-up question, to a need to know even more about how someone goes through their customer journey. We are learning every day.”
Five tips from InShared
Organisations who want to increase their conversion can take advantage of the five tips from Inshared.
- Work according to agile principles, step by step. Try not to do too many things at once.
- Focus on your goal, not on the methods you use to reach said goal. You can spend an eternity thinking about the optimal conversion attribution model, but you will often learn much faster by linking the knowledge gained through profiling to attribution.
- By extension: focus on the areas where you can stand to gain the most. For example, if the average conversion path of your customers consists of two steps, don’t waste too much time on tweaking your attribution model, but use linear attribution instead.
- You can get started without 360-degree customer profiles. Your data is never complete. Just start using the data you do have.
- Make any insights you gain actionable as soon as possible. After all, it’s not an academic session. The goal is to improve conversion. You gain additional insights through implementation.