The concept of self-serve or low-touch has been a popular topic in the customer success community for a while. When we are exploring or implementing low-touch customer success we’re usually trying to address two challenges or opportunities:
- How can we keep the cost of customer acquisition and cost of service low if our ASP or LTV is also low?
- How can we scale customer success to reach a large number of customers by enabling them to find the right resources when they need it and allowing our teams to focus on the customers who need the most human help?
In the first scenario, the need for this approach is critical for your go-to-market strategy. If you sell a low ASP product with a light touch or no-touch sales process you’ll want your CS model to match that. In the second scenario, this is often used as a complement to a traditional CS process in a traditional 1:1 or one 1: few human-led models.
For the purposes of this post, I’m mostly going to deal with the second scenario with an established CS org looking to scale but will borrow from some of the concepts I’ve seen work well in Freemium low-touch teams.
Things to consider -
Before recommending some steps for implementing a low-touch CS programme I wanted to share some guidelines and things to consider to make sure this strategy is right for your team and your business.
Customers should not “feel” low-touch
While the concept of low touch is very appealing at a financial level, not all teams are good fits or ready to implement it. It goes without saying but the most important stakeholder in this strategy is your customer. If your customer is used to receiving attention from a named CSM, moving to a different model should not impact their ability to get their desired outcome or help when they need it. In order for this to be a reality, you need to invest in systems and process that take the operational burden off the customer and make them feel like they have a human on call if needed.
Don’t consider low-touch a pure cost saving initiative -
Related to my first point - to do low-touch and self-serve well you need to invest. Over the long-term, the cost and efficiency benefits can be huge but there will be an initial short-term cost both financially and in human resources. If you approach low-touch purely as a cost saver without the mechanisms to spot these customers are remediate them early you risk long-term retention.
What type of low-touch model are you trying to build?
From my experience, there are broadly three approaches to a low-touch customer experience - where technology is used or relied on to varying degrees
- Tech assisted
- Tech informed
“Tech-driven” customer success is common in freemium businesses or when your customer is small or acquired through a touchless sales process. In this model, automation, community, live chat support, detailed knowledge bas, and docs are used to service customers at scale. Companies like Stripe employ this model and rarely have accounts with a named CSM.
“Tech assisted” customer success is often applied as a complementary service model to traditional CS teams. The most common way to achieve this is through a “team-based” model where your CSMs don’t own customers but share a pool of MRR/ARR and use tools, automation, and self-serve resources to drive success for these customers.
Slack, for example, use this model - they have an Enterprise CS team who work very closely with customers to roll out slack across the wider business but also have low and tech-driven CS to focus on the users or smaller teams that pay for the product. At HubSpot, we have CSM teams who own a book of customers for a higher ASP and complex products but we also have a low-touch team who work with lower ASP and smaller teams/businesses with simpler success needs.
“Tech informed” can just be thought of as traditional customer success supplemented by technology that can prioritize outreach to customers using a combination of alerts, automation, and usage data but with a human CSM being at the core. I wouldn’t consider this a true “low-touch” approach but just an optimized higher tough model.
With all three of these approaches, it’s critical to consider what’s needed and realistic for your team? All require different approaches and timelines to implement.
How is this different from customer support?
What about onboarding and implementation?
Burke Alder has a great post on different engagement models for customer success. The key point in this article that I agree with, is the need to balance high and low touch onboarding with high/low touch post-onboarding. While it might be possible to have a very low touch CS team, the success of these customers is dependent on their onboarding and implementation experience.
If your product requires complex technical setup, change management or extensive training after initial purchase it is often impossible to replace this need with automation or an amazing in-app experience. However if your onboarding experience is exceptional, customers can thrive in a lower touch model.
How do you get out?
If I’m a customer how do I move from a lower touch model into to high touch? There are a couple of instances where this happens more commonly. Firstly, if a customer has a big project or priority in their business that requires more investment and help from your CSM team where do they go? You need to build mechanisms that allow demanding customers to either move back to a higher touch model or (ideally) pay for professional services to reach their goals.
The second scenario which is a core challenge for customer success - sponsorship or decision-maker turnover. I’ve written about what to do when your point of contact leaves before but not in the context of a low-touch model. In this case, these new customers and users might not have been involved in the initial sale, onboarding or implementation phase and may struggle to adopt your product and be successful in a lower touch model without human help. If this is a common challenge for your CS team you will need to invest in ways to surface new users and put programs in place to remediate them.
Must be “as good as a human” rule
If your low-touch success efforts are going to be successful, you need your retention and churn numbers to stay flat at least as you move more customers into this model. There is no point in gaining cost efficiencies if you suffer net retention inefficiencies. Atlassian has a good rule here where they only invest in more automation if it can yield at least, if not better results as a human performing a similar task.
The main point here is that you need predictability in a low-touch model. If you have spikey churn and retention numbers for this segment it’s hard to invest customers into it.
Usage data alone is not enough
A mistake I’ve personally made and heard from other teams rolling out low-touch strategies is over-reliance on usage data to dictate proactive outreach or predict churn. One thing you will miss in a low-touch model is the ability to get a regular “gut-feel” success prediction from CSMs. When this is used with health scoring and usage data you can get a good handle on future risk. The number of accounts under management in a low touch team means this regular interaction is not possible and relying only on usage data can be risky.
With that in mind, you need to look build more qualitative indicators into your health scoring and alerts - how many new users are being added? how often do they need technical support help? Are there any newsworthy changes in their business? Of course, this list depends heavily on your customers and industry.
10 Steps to Implement Your Low Touch Strategy
1) Develop your Customer Success Personas
Customer success at scale requires a deep understanding of your customer or user personas. I’ve written a guide to developing customer success personas here. The main point is that customer personas are very different from buyer personas. They represent the actual users who are in your product each day. Their success depends on their role, profile, level of experience plus a whole host of other factors.
2) Segment your customers by attributes and behavior to determine their appropriate experience
Lincoln Murphy writes about appropriate experience which I think underlies this point. By mapping your personas out you’ll be able to segment your users and understand who can thrive and survive in a tech-assisted model. Some customers are great fits for low touch service while others will not fare so well. What’s more tricky is that this definition can also change over time. A customers service need on day 0 or day 100 may be different from day 150.
The biggest mistake to avoid is simply defining customers good fit customers based only on demographic information. For example “we will service all customers from X industry with Y MRR and Z age in our low touch model”. While this is a great start it misses a lot of nuance in their health, usage patterns, onboarding experience and engagement with self-service resource already.
The ideal scenario is that customers who want to be mostly self-service can raise their hand indirectly by their behavior at any point in their lifecycle. Over time, if you want to get really sophisticated you can develop a predictive model that indicates the likelihood of a customer to be successful in a low touch model based on behavior patterns that link to retention of other similar customers.
3) Test your approach with a smaller low-risk segment of customers
With anything at scale, investing in low touch can be a scary prospect. It can introduce a lot of uncertainty and risk into your planning and forecasting. The question we most often hear is - “if I put x customers into this new service model I may not know if it’s been successful or not for 12 months?”
There are two options to minimize risk here are
i) Ensure you have rock solid leading indicators (see below)
ii) Experiment in advance
Pick a lower value, highly engaged segment of your customer base (this is why segmentation matters!) and enroll them in a component of your low touch strategy. That could be anything from not assigning a CSM, running them through an automated onboarding or only deciding their engagement based on data for a set period of time.
A few tips with this process -
Firstly, it’s critical to ensure you have a control group in place of equally represented customers who are engaged with your standard model to compare with. Secondly, make sure the number of customers you are testing with is significant ensure to ensure any changes in their metrics and success are not subject to chance. Finally, make sure to remove people bias from the experiment. Ensure your teams or anyone involved treats the customers as they would in a normal situation. This will limit the impact of the Hawthorne Effect.
4) Ensure leading indicators are understood and in place
As churn is a lagging indicator and depending on how you structure your billing and contracts, you may not be able to assess the churn and retention impact of your low touch program for anything from 1 month to multiple years.
You need to have leading indicators in place that have been tested over time as they link to long-term success. These indicators can be anything from metrics like team activity, health scores, usage data. Or more transactional triggers like alerts that encourage CSMs to intervene. I’ve written more about leading indicators and goals here.
5) Get detailed about tracking team activity to understand what can be automated
Once your low touch model is up and running you have a massive opportunity to automate and streamline the team and customer experience. Deciding what and how to automate however can be tricky. Ideally, we want our teams to focus on the highest value driving, proactive tasks that drive the most success with customers.
Aspirationally this is great but in reality, with a high volume model, you will have a lot of work that is reactive. Reactive work is unavoidable but it should be a goal to automate or optimize this work where possible. They key here is measurement but this kind of measurement can be tricky. How do we know our teams are working on the highest impact tasks?
One solution is to use some kind of CRM or CS software that allows you to tag every single conversation and customer interaction based on some form of categorization. You could then group these tasks in a matrix of “impact and effort” or even better score each task with effort and impact. You can then use these scores to prioritize what can be automated or turned into a self-serve resource
A very simplified version might look something like below -
6) Get product/operations team buy-in
So you’ve got your personas defined segmentation built, measurement in place and know what needs to be automated - so who is going to build all this great tooling? In most large teams this will fall to an operations team but you may also need to include your engineers to make changes to product, IT teams or other groups.
This all requires time and investment but hopefully, if you’ve run the right experiments and measured what’s working the case will be easy to make. Ideally, before beginning an initiative like this product and operations are brought into the conversation. The general message here is that to do true automation and self-serve at scale it requires a company-wide focus.
7) Invest in customer marketing
Communicating to large amounts of customers at scale cannot solely fall on the customer success team. If you have a customer marketing team this is the perfect group to partner with in your low touch efforts.
Some of the ways customer marketing can help with
- Communication and promotion of things like 1: many webinars, events or user groups
- Communication of product updates,
- Surfacing leads upsell leads for your sales team
8) Build and test automation
Once you have the cross-functional support for your automation and self-service needs it’s time to build out the flows that will have the highest impact. While automation can seem like a holy grail it’s important to tread lightly. Test with some low-risk segments or actions first and scale up over time.
9) Define CSM roles and responsibilities
What is the function of a CSM in a low-touch team?
Similar to a traditional model there will be a mix of proactive and reactive work. As we identified already, this role should not be 100% reactive as it will simply become a support function. It’s important to set goals and thresholds early of how much reactive, proactive and program work team members should be involved in.
For example, maybe your team spends ⅓ of their time on reactive inbound work, ⅓ of their time proactively engaging with a segment of your customers to drive success and ensure renewal and ⅓ of their time partnering with marketing, product or other roles to gain efficiency and scale for the customer base.
10) Test and tweak your communication channels
Success in a low touch model is determined by your ability to reach a large number of customers with efficient communication channels. The traditional channels - email, phone, face to face and QBRs need to be replaced by channels that can reach many customers at once.
Every audience and market responds differently to different communication channels. Webinars might work well for one type of customer while forums or message channels may be better for others. My advice is to test analyze and test again. Here are some ideas on channels to test out based on their potential reach and effort -
I hope this guide gives you some thoughts and structure for defining your own low touch journey. This is a rapidly evolving area and still very new to me so there’s a lot learn. If you have experience or advice of your own to share please let me know about it in the comments or get in touch with me directly.