13 Best Tactics for Creating SaaS NPS & In-App Surveys (2025)

💬 Running surveys (in-app or via email automation) for NPS and other qualitative/quantitative feedback needs, is an important initiative that SaaS growth marketers, PMs, and founders need to do in order to close feedback loops to help with product improvements, prioritisation of what to work on, detect reasons why certain users/segments may churn, and numerous other needs.

And whilst you've got your core foundational reasons and the "why" of needing to run surveys, how effective they can be for various objectives around feedback that you can action is where the real work comes into it - and from over the years doing marketing within SaaS/B2B, surveys/data feedback plays a big role, and I've certainly got some learnings/tips that I recommend, and things to keep in mind.

Different lessons, considerations and tests to run for improving response rates and data collection quality for SaaS

💬 With some of the checklist points below, I've made sure to include some software recommendations that's worth looking into to help get some initiatives off the ground faster.

This is one of the first checklist points due to the fact of it's importance in my opinion - there are numerous software vendors out there which are great and help with everything from analytics, segmentation, design, etc - and most importantly, provide great UX experiences for users (and also for marketing/product managers).

Plus, the opportunity cost of setting this stuff up custom just takes too much time (and I can say that from experience).

In order to increase NPS/in-app survey participation, as well as ensure the right data is collected (which you can action), you need to ensure to triggers prompts to user segments that makes contextual sense.

So, you need to make sure:

• Trigger surveys based on the CX journey - when/where you trigger and what you say needs to be tailored - e.g. free-tier users are different to paid 'power users'
• To ask context-specific questions - e.g. questions you ask freemium users vs 6 month paying users are very different (plus, the data you want to know is different - so you need to cater)
• Tailor follow-ups based on the surveys you're doing for each segment/stage

One of the more important checklist points to bring up here (and related to the above) - you need to ensure you've planned out the segments and triggers.

Then, it's about ensuring to 'map' the different types of in-app NPS prompts and surveys to get the right data you need to action things.

So, what are the various types of NPS/surveys that you can do? Here are some that I've used over the years:

• Feature adoption surveys (which I mention more about about three points down)
• Churn prediction (which I touch on below)
• Post-onboarding - display within X time periods (hours or days even, depending on your offering/nature of product) - this is also a good time to ask questions based on user attributes
• Milestone-based - trigger based on timings or certain behaviour/activity levels (e.g. 14 days after signing up, etc)
• Community-driven and/or educational content - questions and surveys related to community and/or content
• Customer support - feedback for support (which many SaaS brands do)

And the list goes on really in terms of types you can trial.

Related to the above and also just as an important checklist point - collecting survey/NPS data is one thing, but the real value comes from what you can do as a marketing/product team is provide help contextual information faster - i.e. close the feedback loop to provide help solutions.

This is where setting up conditional/dynamic path follow-up sequences come into play.

As a high-level example of what I mean by this, say you want to run an NPS to recent user signups. You want to trigger when users have signed up 30 days ago, and want to ask a "scale" question, where users can pick any number from 1-10 (lowest to highest).

From there to give the right paths based on the answer, you want to have conditional paths based on these answers - here's a basic layout:

• 0-4 - in most cases, this means that users are not receiving great answers (i.e. low satisfaction) - so, you ideally want to have a conditional sequence focused on routing them to customer support.
• 5-8 - typically means they are somewhat satisfied - meaning you then need to have a mix of customer support and resources that you can share.
• 9+ - then for answers 9 and above, you can then potentially ask for a review on a 3rd-party platform.

Note - this is purely a high-level example, but you can see what I'm conveying - context and ensuring the right personalisation is there to help support based on what segments/users answer is how you can effectively use NPS.

This is more of a "type" that you should have setup - definitely send in-app prompts to users that seem to be disengaged, have dropped off in terms of usage, or other signs/signals that matter within the context of what your product offers.

Focus on potential pain-point questions and make sure to prioritise follow-up flows/sequences to address at-risk segments.

Whenever you release a new feature, whether closed to beta segments or you've wide to all segments, you want to try gather feedback fast - send out prompts about the specific feature and keep question types very simple (i.e. '1-click).

Based on the conditions of questions you setup, you can tailor follow-ups from there and setup internal alerts for customer support, etc

As a general rule from experience I've found when doing data collection/feedback campaigns, I typically focus on keeping prompts micro.

You don't want the user to overthink or be distracted from their experience, as well at the same time, you want to avoid dropoff.

You just want to get 'standardised' data points quickly, and from there, you can then use data and trends to create customised follow-up sequences (which is where the value for users and your goals come into fruition).

This checklist point is a bit of a no brainer and aligns with the others I've touched on - make sure to align messaging from the initial pop-ups within app, all the way through the automation flows.

Messaging as to make contextual sense to the user/segment - otherwise, you simply won't get the desired results you're wanting.

As something worth testing and depending on the CX journey of the segment, test qualitative questions after asking for a simple '1-click' question (True/False, multiple choice, etc).

You can of course do this based on conditions of the '1-click' answer that users select.

Related to conditional paths, you want to ensure you have a good follow-up sequence and path for users to help solve problems they may be facing, or how they can unlock even more value from your product.

For instance, here are some example cases you can consider:

• Those who aren't satisfied (those who chose low scores) - you could share product update emails/posts focusing on requested features (to show you are listening to your users/customers), share case studies/social proof from happy customers (and ideally target based on related ICP attributes), build feature spotlights addressing passive user/customer concerns, etc
• For those who have shown satisfaction and loving the product - you could ask them to record quick win videos (which you can use for UGC type ads and social posts - which you could then use for campaigns for inactive/less engaged audience segments), feature their testimonials/positive words in relevant product areas (in-app), create solution-focused blog posts, and more.

Plenty of ideas/angles of content creation you can do, which can help with development of automated flows/paths. :)

I certainly encourage to trial and change up the question formats, but also, it's worth testing and experimenting the formats of questions - so that being single questions, emoticons, or even the likes of image-based.

I also encourage to switch up questions to keep users engaged - it's good for testing, but also good for 'pattern interrupt' - simply meaning that it changes expectations of what many users would typically expect from a SaaS/product asking for feedback.

Here are some key reasons for some of the questions formats to keep in mind when testing:

• Rating scales (typically the '0-10 slider' you see) - I like using this format as it's good UX for users as they intuitively understand it, and the visual slider makes it feel more interactive - plus, it helps make it clear for benchmarking and tracking
• Multiple Choice (which we all know quite well) - I like multiple choice for categorising common issues and is good for making sure people do answer - I also recommend to include an "Other" option with text field for those wanting to share more context (don't have to make compulsory, but good to have there and I've found from experience that quite a few do share).
• Emoji reactions - this is a good format to gauge emotional-based responses - I've found from experience that they have higher engagement rates, and feels a little less formal and at the same time still provides a key data point to track
• Consider picture selection - don't always see these, but have come across a few and I quite like it, especially around feature requests/feedback - good chance will help with engagement/response rates

I've already talked about testing/experimenting across various types (questions, triggers, etc) - but testing also comes down to other elements, including design, types of messaging, CTAs, and even channels and flows like email, and within the product.

Depending on the software vendor you use, you should also elect 'A/B split-test' options to test variants.

And of course, one of the more critical factors of running this initiative is reporting - and you want to have reliable reporting based on the multiple NPS and surveys you're running.

This is where choosing the right software vendor really can make a difference - and one thing to look out for are AI capabilities to help dissect and get important insights and information faster.

This can help with multiple areas for future data collection/feedback campaigns you run.