We sent out the offer letter for our marketer today, and we are ramping up our marketing efforts for Nubela. Since I started writing daily, I have subscribed to a fair bit of transparency. And in the same vein, I decided I will "open source" our step by step playbook. Also because I am lazy and I have decided to qualify my day to day specification-writing work as a blog post.
Following up from yesterday's post on "Total addressable market", the very first task for our new marketer is to identify our target audience. Here is the task specification if you are interested:
Research target audience
This is a research task. And the very first task for marketing is in identifying who our audience is. Thereafter, whatever we do, we can personalize our efforts, be it content, tone, or engagement specifically for this niche.
Definition of the ideal target audience
Our ideal target audience is an industry for which we have the perfect solution that will solve a grave problem for them.
Examples of a perfect niche
- When Grab first launched, their biggest problem was getting enough drivers. So their initial target audience was drivers (and taxi drivers). They had people situated at car parks / coffee shops (for taxi uncles).
- When Gom VPN launched, I targeted men who obviously watch porn. Do you know where they can be found? Hardwarezone EDMW forums.
Identifying our target audience
To identify our target audience, we have to know who we have to sell to. To figure out who we are selling to, we have to understand what we are selling.
The problem is, we are selling quite a fair bit of things. But this is normal. Most companies have a myriad of solutions for a scope of problems. Instead of looking at the solutions that we offer, break down our solutions into the tools that we have, that we provide. To simplify it:
- We have Proxycurl - a tool for developers to get (private) data from the internet to build up their form of a dataset. -- In other words, we have a way to help businesses procure data, and also to build our dataset.
- We have Facial Recognition Technology --, we can identify people from pictures.
- We have a team of humans -- When we do not have the perfect solution, we can leverage our human resources to customize a solution for the customer given the above two tools.
These are the fundamental pieces that make our capabilities.
Not a Singapore business
Nubela is an internet startup. What that means is that we should never limit our audience to Singapore only.
The exception being hyper-focusing on a local niche for pronounced effect on marketing. For example, we decided to target "AI companies." Instead of targeting AI companies globally (there is a lot), we will produce a subset of our marketing content to be focused on SINGAPORE AI companies, which narrows a wide niche further.
The smaller the group, the easier it is to personalize the content that tug at them.
Focus on a segment for which we can deliver now, which is Proxycurl + Humans. Our facial recognition/Sapiengraph demo can only come to life in the next two months.
- A report written for this task on Giki on this page: (REDACTED)
- In the report, it will be good if you can list and consider various market segments including but not limited to
Analytics companies (Consulting companies like Bain/Mckinsey/Deloitte need data)
Other data companies
- Considerations should include ease of access to this segment (is there a forum/community by which this segment is heavily represented in it?)
- Considerations should include the size of the segment.
- Considerations should include how grave this segment needs a solution that we can uniquely provide.
- A draft to be submitted at the end of Monday, and end of Tuesday so that Steven can review it.
We do not have deadlines. But let's scope this task to be presented on Wednesday morning (24th July 2019) at 2 pm. If you are unable to finish it, tell me so I can reduce the scope of the task.
Resources for further reading
- Look at the solutions and customers our competitors in the facial recognition space have: https://giki.wiki/@nubelacorp/sapiengraph
- Look at Diffbot, our customer who pays us for Proxycurl to build their own AI dataset: https://diffbot.com