Explore our Digital Experience Management Toolkit

7 Customer retention techniques to increase your ROI

Author's avatar By Expert commentator 17 Apr, 2017
Advanced Advanced topic

The 7 elements you need for superior retention automation

It costs a lot to acquire a new customer, but most clients will make a single purchase and leave. A repeating purchase, on the other hand, costs far less than acquiring a new customer, estimated at between 4 and 7 times less actually.

Retention automation has a simple, well-known theory that deals with this issue. Retention automation aims to get better ROI by turning one-time buyers into repeat customers and keeping them, rather than focusing on acquiring additional new customers.

Retention automation can be done with the basic understanding that customers are different, then analyzing them as groups, connecting them to automated campaigns designed to move them in the right direction.

There are plenty of examples that demonstrate programs and multichannel campaigns addressing customer retention according to the customer’s current lifecycle situation and their importance to the business. Surprisingly, although the theory is so simple, most companies aren’t doing it - the result of a combination of obstacles.

7 steps for customer retention

We believe that 7 elements can go a long way in resolving these issues.


1. Automate and simplify data collection

Data centralization and big data models have been broadly discussed lately. Still it may seem quite a challenging technical obstacle to overcome for some companies that can’t invest the IT resource to connect all the different data silos together. The retention automation platform does this, reducing the likelihood of marketing becoming another needy customer for IT’s attention.

Specific data should be used for specific purposes, and retention automation is quite specific so it should be able to serve itself. Therefore there are 2 ways a good retention automation platform should help you succeed in this fundamental part: first, it defines a simplified data model to include only what is needed, not to cripple the entire project. Second, it collects as much as possible automatically for you so you don’t need to bother with it. Let’s elaborate on these two points just a bit.

Simplified Data Model
The simplified data model: the very basics are your contacts, their purchases and the products they purchased. With these 3 tables alone you can already do quite a bit, for example, this is sufficient to allow you to do the basic RFM (Recency Frequency Monetary) segmentation. Your email behavior, your website behavior, your app behavior , must all be added to include the engagement data, as this defines what your contacts are doing now, and will do in the future.

Automated Data Collection
The automated data collection: meaning using scripts implemented to your website without disturbing it, or integrated automatically to your shop database so the data flows automatically. When scripts are not desired, the retention platform should provide a very easy and secure data transfer protocols for the daily files.

2. Built in scoring and segmentation

Now that the retention automation platform has its data in the designated and focused structure, it will use it’s built in knowledge and predictive scoring models to build smart and effective segments for you. Distributing all your contacts into clustered groups that makes sense to be treated separately.

The very basic level of this is RFM analysis. If you add engagement scoring into the mix, you get ERFM. ERFM may appear to be simple (and you can read more about it here), but it harbors more options then you are likely to use, for example: most companies don’t have this in place one crucial program that automatically wins back defecting buyers, with an adequate incentive according to their buyer status and spend.

After you cover and have used the basic customer lifecycle segments, scoring and segmentation can go further, into micro-segments, cohort groups, predictive clustering etc., to become even more targeted. The sky is the limit really and a retention automation platform could take you there. But we recommend starting with the basics before tackling too many advanced frontiers.

3. Smart metrics (analysis powered by predictive analytics)

Let’s look deeper into our segments. Each one of these customer lifecycle segments is living and breathing since it is made out of people. Obviously you should know them, but how can you get to know 100,000 people in your defecting buyer segment personally? Well, you shouldn’t. You can’t (and shouldn’t) care what each one of them bought and how long he stayed on your website because you don’t need to manage the singular details.

However, applying a kind of x-ray vision to the mass allows you to look at the segment and discover its aggregated personality, which highlights the key metrics that are fundamental factors for your success.

From data to insight

Here is an example: did you know that if your first time buyers did not make a second purchase after 47 days, they are more likely to defect rather than buy a second time? You guessed you had 90 days, but how much it was based on real time data?

And even if it was true once, this is a breathing, evolving segment of 100,000 people, it is bound to continue changing. Look at the trend of this change over time, not only because you like statistics, but because it will make you change your marketing program. In our previous example, the incentive to make the second purchase should be changed to meet the client at the 42 day mark, and not 85 days.

Predictive behavior modeling

Another example of smart metrics is with built in predictive modelling. This allows you to see that if you continue do what you do now a defecting buyer is 7% likely to convert, and an inactive buyer is 0.08% likely to convert. This will highlight when it is time for a change in strategy for these segments.

Motivational metrics

Another type of smart metric is motivational metrics, i.e., reporting that motivates you to act. What is the ratio of first time purchases in your business? Did you know that most money comes from first time buyers that will not buy again? Usually only 30% of revenue comes from repeating customers, if you are above it you are doing well compared to the industry, but this number should be higher, across the board – can we really accept that most customers will never come back? We offer a report that shows you exactly that and on a trend saying what is the impact of retention automation for you and are you getting better at it?

4. Predictive revenue

Smart metrics from the field of predictive revenue calculations go deep and tell you how much money each one of your defecting buyers is actually worth for you now, how much is a cold lead worth: how much revenue you will generate for the relationship lifetime by converting, say 245 first time buyers yesterday, and how much money you probably lost by missing on the 652 defecting ones that yesterday became inactive.

These are only predictions of course, but they come pretty close, as they are based on your own data. Provided in a single screen dashboard, you get a bird’s eye view of the revenue impact for your actions (or lack of actions), and you see potential, and the wider picture starts to make sense.

In our platform we aim to characterize each segment with 10-20 smart, meaningful metrics like those mentioned here, that really tell the story, help you automate your campaigns, monitor trends, and reach decisions about your marketing.

5. Integral part of your engagement platform

Here is a barrier that is easily underestimated: you might have your own business intelligence solution in house, at whatever level of sophistication, or maybe a niche solution focused on a specific area, and it may be integrated to your engagement platform in some level. But if it is really a part of it, things get lost.

It must be part of your core marketing platform that handles your contacts, your segments, your automation, your reporting, your content, etc. otherwise it will be slow. It will be inaccurate. And it will cost a lot of money.

At the end, you will do much less with it. Marketers should carefully challenge their vendor strategy and aspire to have everything in one interface, or risk blocking opportunities. You should be able to see, analyze segment, create content, automate and report in one user journey.

Integration includes triggers, meaning if a customer did something on your website, an action would be triggered to the platform in a seamless manner. Content also should be fully integrated and smart. It is not the focus of this article, but content is of course a critical driver of successful campaigns. Ultimately, every customer should get a message that really is relevant.

To deliver that you need good campaign design, good buyer status treatment, and of course good products. But your products need to be better than good , they should be right products for that person. This 1 to 1 product matching at scale can only be done with a recommendation engine. Retention automation platforms that don’t automate this kind of content will leave you short of one of the fundamental ingredients in the mix.

Integration also means access to multiple marketing channels. While email is the most reliable and tested channel, it can’t be the only one. If the email didn’t work for the defecting Platinum buyer, send them an SMS - you must do whatever it takes to win them back

Measuring impact

Imagine you have tailored your strategy to include the new capabilities described above, and suspect you’re engaging customers better than ever. Can you measure this impact? For example: you are setting up an automated campaign designed to win back defecting buyers. You include a nice incentive, great product recommendation, you turn it on and watch what happens. After 3 weeks, you won back 212 buyers. Is this good? Better than nothing! But could it have been better? Maybe. Who knows? If you had done nothing, would these 212 defecting buyers come back anyway? Or maybe only some of them?

Looking at opens, clicks or even conversions won’t do the trick. The only scientific way to be sure your campaign made an impact is to have control group testing. If that program was sent to 90% of your defecting buyers, and 10% of them did not receive it, you would be able to see how many of the 10% came back anyway without you approaching them with that campaign and extrapolate the value of your strategy. You’d know how much the program is worth. Meaning not just revenue generated when you sent the campaign, but revenue that can be specifically attributed to that program as it would not have come otherwise.

6. Meaningful reporting

Data can paralyze you. The bigger the badder. Even with visualization.. Retention automation is not a full business intelligence solution, and it shouldn’t be. For the sake of retention automation marketing the approach should be a specialized system that collects only what is necessary to do its magic and show you only what you need. Otherwise, metaphorically speaking you can’t see the forest from the trees.

Preplanned reporting templates, designed by data scientists, focused on retention automation will clarify the next best actions. This is reporting that counts, that visualizes exactly what you should be looking at, whether it is your segments, your performance, the smart metrics, it is tailored to your needs and empowers you to make decisions based on knowledge.

Implement an efficient daily reporting routine: manage, motivate and monitor.

  • First, manage your assets: these are your contacts. See how they are distributed across lifecycle segments, learn their aggregated personality, and how it changes over time, allowing you to analyze them with the metrics that count.
  • Second, (financial) motivation: seeing where the potential is. This problem costs this much money, so focus on this goldmine right here. How much of my revenue is from first time buyers, what is the ratio of my refunds, etc.?
  • Third, impact: monitor the results of the actions you are taking, what preforms well, and what needs a change.

7. Encouraging experimentation

People tend to be scared of change. Companies are typically very set in their ways. It’s not always intuitive to try new things, or to change something you think already works well. You would consider a change, if you could see that it would lead to improvement. Well, the same control group testing concept discussed before can be used in this case.

Allow 95% of a segment to continue as before and divert only 5% to a new concept. Your webmasters should be doing that all the time online, and you should also make sure it is done in campaigns and automated programs. Changing segments, content, incentives, channels, timing - anything goes, let the results speak for themselves. Experiment didn’t work? Fine, turn it off. If it worked and looks better than the original – expand it to 95% and you are ready to try out the next one.

Make it happen
By now many marketers have got their hands on the big data models, but the problem still stands. The adoption of these obvious ideas is not rising. Coming from the school of product management we like to think that the solution to all problems lie in software. We think of what we can do with the software so the user will end up using the technology - the answer is what led us to the 7 elements described above:

  • 1. Automated and simplified data collection
  • 2. Built in scoring and segmentation
  • 3. Smart metrics (analysis powered by predictive analytics)
  • 4. Integration with your engagement platform
  • 5. Measurement of the impact
  • 6. Meaningful reporting
  • 7. Encouraging experimentation

These give you what you need to achieve retention automation, and will motivate you to do so.

If you do, you will dramatically improve your bottom line by maximizing total revenue generated from existing and loyal customers. Use these 7 elements as a checklist when you evaluate any potential retention automation solution and you won’t go wrong.

Author's avatar

By Expert commentator

This is a post we've invited from a digital marketing specialist who has agreed to share their expertise, opinions and case studies. Their details are given at the end of the article.

This blog post has been tagged with:

Predictive analytics

Recommended Blog Posts