Learnings from launch: Introducing powder dispensing technology into informal markets

Photo credit: Eugene Kaiga

The following is a guest post written by Zahid Mitha from Novek, a company that builds Internet of Things technology for dispensers and vending machines for micro-retailers.

Over the past six months, Novek has been developing a dispensing system for washing powder in Kenya. This project is part of a partnership with Strive Community to address the high financial and environmental costs associated with the packaging of this regularly consumed household product.

Novek believes that creating and piloting dispensing technology in Kenya could provide the data that will allow us to scale globally and positively impact micro-retailer sales worldwide. Micro-retailers in Kenya exhibit traits common to other small store owners in many emerging markets: they sell products in varying (and sometimes very small) quantities to consumers on a huge range of budgets, and they do so in informal settlements. Thus data from Kenya can support the development of similar technologies in other contexts. In a previous post, we shared our insights into the needs of these informal retailers. Based on these insights, Novek has created and launched our first prototype into the Kenyan market. This post details the launch process, including highs and lows so far, and discusses what we’ve learned to date.

What did we do to find a retailer that would take the machine?

Finding a micro-retailer to pilot our machine proved to be a challenge due to several factors:

1. Low trust among informal retailers: Many small retailers do not have owners present on-site, or they have low levels of digital literacy — both of which lead to lower trust among small retailers. There was a need to educate them about the product, how it would benefit them, and how easy it would be to use. Fortunately, we had laid the groundwork with previous surveys that informed us about the needs of these micro-retailers, their potential concerns, and their objections, and also gave us a pipeline of possible retailers who would want to use the machine.
2. Space needed to be measured in-store: While space was not a constraint due to the fact that we had used survey data to adapt our product to informal retailer environments, we still had to go and confirm suitability on-site before choosing a retailer, including ensuring they had a tabletop. This added time to our process. A further unanticipated challenge was that some of the retailers we had surveyed previously had moved, closed, or changed ownership. (This is, unfortunately, a necessary evil given the lengthy product development cycle for a physical prototype.)
3. Novelty of our machine: This machine is the first of its kind. While explaining the product conceptually was relatively straightforward, in practice we had to surmount digital literacy barriers. We spent significant time teaching the retailer how the technology worked, how to interact with it, and how doing so would benefit their business.
4. Accurate sales data: We found that, while the previously stated volumes of washing powder sold by some stores appeared to meet our criteria, in practice, this was not the case either due to misinformation or lack of recordkeeping. As collecting transaction information to assess pilot success was key, we had to move the unit when a location proved to sell less detergent than it had reported.

Despite these challenges, we were able to find our first pilot retailer in a relatively short span of time. After shortlisting a group of potential sites and making both a qualitative and technical assessment of them, we selected one for our first prototype. This micro-retailer was ideal in several ways: the owner was receptive, the store was sufficiently high volume, and the owner was open to giving us regular feedback.

What was the retailer’s experience like in the first two weeks?

Our first week was rocky, to say the least. As soon as we placed the unit in our selected store, the shop experienced three days of ongoing power outages, resulting in the machine not being used at all. We have had experience with power outages before, but they tend to be relatively rare in the areas we target.

There were further delays, as the shop had run out of washing powder. Finally, the shop owner decided to completely relocate the shop to a more profitable area. Assured that the new site was still secure and met our criteria, we deployed with them.

Halfway through the second week, the new shop location was up and running, our machine was stable and stocked, and we had run training demonstrations at the shop with the owner and his employee.

Novek’s washing powder dispensing machine in a micro-retailer’s Kenyan shop. Photo credit: Eugene Kaiga.

There were still further challenges to come, however. Given the complete novelty of the machine in the market, customers were curious about how it could help them. However, the owner’s employee wasn’t able to fully articulate its benefits due to initial unfamiliarity. We worked hard to train them on these benefits for customers, and we’re already seeing this strategy bear fruit.

What were our initial learnings?

We’ve started to see some early success. Sales are growing at the store, which is close to dispensing 3 kg/week of washing powder (up from zero just two weeks prior).

We’ve also learned a huge amount from even this short period about how we can further adapt our product to the needs of micro-retailers:

1. First, pricing the product was a complication for the store. They were used to selling smaller packages of washing powder. With our technology, they instead fill the machine with larger quantities which could be dispensed. Retailers found it hard to determine the correct price per gram sold. We will continue to work with them on this and believe a long-term solution will be achieved through future partnerships with manufacturers, in which pricing is pre-set for bulk purchases.
2. Second, we’ve learned how to make the user experience easier for small retailers to quickly grasp. This includes everything from price setting to dispensing and loading. We are rolling out a new version of our software with these learnings over the next few weeks.
3. Third, we have realized that product supply interruptions, while partially out of our control, are key to ensuring regular trust and usage. We are working with supply partners to minimize disruption.
4. Fourth, we have found that the machine may be inaccurate due to the unpredictability of retail conditions if it is not calibrated. This means that the machine dispenses unreliable quantities of product. We are working on both technical solutions to improve accuracy and retailer education to ensure smooth usage.

Next steps

Our upcoming plans for the project involve a month-long testing phase for the machine. This will enable us to learn more about how the product is being used, make technical improvements and tweaks, and collect data that will help us better understand how we could eventually scale.

After this period, we plan to broaden the product’s reach to three more retailers in the same target market. Ultimately, we will collaborate with business partners to launch a commercial pilot of the project. Watch this space!