More with less: Implementing data-driven decision-making at Tamara

Tamara’s product culture has been built on a few core pillars namely:

• Customer focus (solving a real customer pain)
• Perseverance i.e. relentless pursuit of making things work (“finding a way” is one of our core values)
• Data orientation

In this article, I will focus on how we work with data and what makes implementing data-inspired decision-making uniquely challenging for innovative start-ups like ours. This article is NOT exactly about how you use data in the software development life cycle, but more about how you supplement the lack of data and STILL remain true to the tenets of being data-driven in a fast-paced innovative setup!

Product managers at Tamara use data extensively during all stages of the product life cycle. What makes utilizing data for an innovative and disruptive product like ours interesting is that there is quite often not significant enough data to derive insights from and there is sometimes zero data to start with.

So how do we break this apparent deadlock?

Resourcefulness and scrappiness: Working with what you have!

This is the fundamental guiding principle we apply at Tamara. While one can derive additional insights and probably clearer actionable inputs from more data, sometimes you just don’t have it or, I would wager, you don’t need it!

So it’s better to look at what other data points would approximate insights you are looking for and then work towards gathering those data points. In general, please remember if you don’t have access to all the data points and you likely need to move forward, it’s a good sign! Otherwise, your product is probably too late to be released to the market.

Here are 6 specific pieces of advice to operate effectively when you don’t have all the data you need:

1.Getting comfortable in operating with insufficient data. The first starting point is acknowledging that if you are innovating or solving a problem without much precedence, data availability will be a constraint.

2.Start with industry-specific data points. Let’s say you are building a content product wherein you want to gauge what kind of content your customers like. For such scenarios, looking at Google Trends, Facebook’s Audience Manager, Google Keyword Planner, and site analytics from Google Webmaster are excellent sources to approximate customer demand. However, be aware to filter out data for your target audience based on geography, demographics, and other behavioral nuances.

3.Pivot to competitor / industry benchmarking. Products at Tamara are built to exceed customer expectations–which in most cases are defined by what they have experienced in other contexts and markets. Therefore, product managers at Tamara focus on understanding how the industry benchmarks around specific customer solutions look like and then work to create a WOW experience for our customers!

4.Look at the secondary data sources internally. Have an objective assessment of what data points will correlate to your use case and try to find the closest proxy towards that objective. When we were launching the Buyer Protection Program at Tamara, the team looked at data around customer tickets to prioritize roadmap items directionally. Since previously the program didn’t exist in Tamara, it was not possible to use quantitative data for prioritization.

Internal site search is another excellent source to understand what customers have been looking for. As an example, you can use this data point to estimate how customers view you and the competition. You would be surprised how often people would search for your competition on your own platforms!

If you are classifying customer intent and interest based on query and volume analysis, be mindful of the following biases:
a. Some customers may struggle to articulate their exact needs.

b. For innovative products, a low volume doesn’t necessarily indicate a lack of demand!

5.Evaluate alternate types of data. Sometimes when you don’t have access to quantitative data, it is perfectly okay to gather qualitative data and understand the customer perspective around problem space. Now the question arises, how many customers do you need to talk to? The exact answer of course depends on your customer profile and the problem but I have seen customer insights tend to converge around 15–20 in-depth customer interviews. If you have the prototypes ready, 6–8 customer interviews should also be okay.

6.Operating with a long-term mindset by investing in gathering data is key. One should be willing to invest in generating primary data insights from the customers. If that requires investing in specific products, tools, or technology, it should be okay. There are many plug-and-play tools one can integrate to understand customer journeys, create heat maps for critical pages, understand the most unhappy customer scenarios by analyzing the last event behavior before drop-off events, and even explicitly ask customers about their experience.

For some of the tools, be extra vigilant that you are absolutely compliant with Play/Google Store’s Privacy Policies and of course with country-specific regulations you operate under. When in doubt, please disclose to your users how you are going to use specific data and any extra permissions you may need from them! In such scenarios it’s a good practice to disclose proactively how you would NOT use that feature itself

Let us know what other methods, tools, and tactics you have found useful, we’d be glad to hear back from you!