Data is the new Oil — but how to use this oil for more profits?

AnkitG
AnkitG
Aug 24, 2017 · 4 min read

I am sure almost everyone has heard and read about this statement — ‘Data is the new Oil’ But what do we the big companies do once they have the oil and a lots of it, what are the possible by products? How can we use the oil and make petrochemical products like stitching yarn?

One of the challenge for the big companies is what more can be done with the data like customer behavioural data, sales data, UI interactions, cross device data, location data, beacon ping data and so on.

Most companies use it to make better reports and visualisation — yes we have tools like Tableau, Answer Rocket, Qlikview for this but what else?

I have worked with companies who had data for 3M- 40 M (yes that is equal to half of German population) customers and were recording billions of interactions on their apps every day.

But how can you utilise this huge volume of data? As a product lead for data products and working closely with data scientists this is one question which I am asked quite often by the business leaders.

Well, take inspiration from Oil economies, after struggling with oil prices they realised that they have to find new avenues of making money and need to diversify and create related industries like investment banking, precious metals and real estate and so on.

Also, my earlier work experience in Research and analytics came really handy for me in answering this question.

For any research problem we used to take explicit user feedback, conduct interview, focus group discussion and sometimes also a market survey. Obviously all these approach have a major problem of limited sample size.

So, the answer is, use your data to build products for research and analysis and may be sell this knowledge to other companies and create an alternate income stream.

The advantage of having a lots of customer data (~20M customers) is that we can run analysis on the whole population and not just the sample. ( this can be argued because still limited users are online and from them not all will visit your website)

Now, these days there are easy and cheap solutions to capture and store humongous amount of data (s3, Google cloud, Azure) and with the help of big data technologies (NoSQL, MongoDB, ProsgreSQL..) to mine this, we can overcome such challenges. We can build internal market research tools which can answer precisely what the customers want — not just any customers but only your specific target group.

At my previous company, we built an in-house market research solution and called it Customer Insights platform to help category managers understand our customers better at a very micro level.

We were also able to help brands understand their customers better and their preferences for other related categories and products.

Thats not all once you know your target group better you can actually uniquely identify them and use it for marketing campaigns across any of the communication channels (including Facebook for look alike campaigns)

We built this platform by aggregating internal transactional data of more than 3M customers and behavioural data of > 20M visitors. Users can slide and dice customer profiles using more than 50 dimensions, and can serve answers to more than 50,000 permutations and combinations.

The main categories were -

Product taste (brand preferences, colours, fabric, pattern..)

Demographics (age, gender, city, state, ..)

Purchase data (selling price, discounts, coupons, date of transaction, number of products, wardrobe penetration ..)

and two special variables to create better customer cohorts

Activity percentile (weighted average of engagement with our apps, includes recency, frequency of usage)

Affinity score (affinity of the selected cohort in comparison to all the other users which are part of this product platform — CIP)

I am sharing some of the sample research questions which the tool can answer for the business owners.

It took us not more than 3 months to build this product which has enormous capabilities.

eg.

  1. Cohorts of two different cities, can be built and compared.

2. Based on objective of up-sell or cross sell we can identify the right target group (age, gender, city..) and then design the creatives this cohort of users will like (color, language...)

3. One particular cohort of users who have a preference for one brand can help the brand manager understand what other brands his target group seek (Nike vs Puma, vs Adidas vs Brooks vs Solomon..)

4. For any target group, we have customer identifier and and using that identifier (email, Google advertising id, UUID.. ) we can actually execute personalised marketing campaigns for their taste.

Hope this post enlightens you with the possibilities of using the data the new oil for building better data products.

Do share with me, how you used the data to build new products and then also create alternate revenue stream from them for the company.

Look forward to your comments and suggestions.

ps — In the next post, will share the technical architecture of this platform.

)

Written by

AnkitG

Rookie photographer & traveller. Professionally I enjoy building data products which lead to improvement in customer engagement & retention.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade