How to start your own data analytics startup
HTML-код
- Опубликовано: 21 дек 2020
- My book Data Culture (amzn.to/44Fe1cu) is out.
‐-------‐‐-----------‐------------------------------------------------------------------------
After a while as a data analyst you may decide that you want to start your own analytics consultancy or agency.
This is different to becoming a freelancer, it's doing something that you can grow and hire others who can help you scale.
Based on my experience of creating an analytics team within a larger agency and having also started 2 analytics consultancy start-ups, I share on this video what I think are key to making it a success.
Here is the link to the lunchclub website
lunchclub.com/?invite_code=sh...
----------------------------------------------------------------------------------------
Please subscribe to my channel.
If you have any questions please do put them in the comments below and I will answer them in the comments or may create a video to answer your question.
Like the video if you enjoyed watching, it helps me understand if the content has been useful and relevant.
----------------------------------------------------------------------------------------
Subatomic Analytics - www.subatomicanalytics.com
I currently run my own analytcis consultancy, so if you would like help or advice on anything related to data and analytics please do reach out to me at shorful@subatomicanalytics.com
----------------------------------------------------------------------------------------
To know more about me, check out my LinkedIn profile: / shorful
This is my Twitter account: / shorful
----------------------------------------------------------------------------------------
The views, opinions and position in this video (and other videos on my RUclips channel) are my own and no way reflect the views, opinions or position of any company I have worked for, or worked with, in the past or currently.
![Chk Chk Boom - Stray Kids ストレイキッズ 스트레이 키즈 [Music Bank] | KBS WORLD TV 240719](http://i.ytimg.com/vi/63QPwqCRGcw/mqdefault.jpg)
I just discovered your channel this week, and I have to say you are extremely underrated and due to go big! Your content is so rich, but distilled, and your delivery is lovely.
Thanks for the kind words.
I've not been able to commit much time to this channel due to running my companies, but I'm hoping to be more consistent with the videos.
If there are any topics you want me to do a video on, then let me know.
Please subscribe, keep watching and share my videos
@@Shorful I second this comment, just from this video I would watch a series on Data like this. The way you explain concisely is underrated and packed with value
Thanks Patrick
Really appreciate the feedback
Very underrated channel. Thanks for the great advice.
Thank you for your kind words
Please do continue to watch and let others know about my channel
Your amazing thanks for this!
I'm still going through the IBM courses and really excited to do something like this. I know it will take time but, it's still interesting to me to start this type of business.
Keep going
It's definitely a good business to do
I'm busy with the same process. Just in a matter of time
Good luck
Excellent video! Thanks for sharing your experience
Thanks, hopefully can do more of these videos but been busy with running my companies
Thanks for the great knowledge, and guidance.
My pleasure!
Any other data terms you want me to explain just let me know
Thank you for sharing!
This is great advice, thank you! 👍🏼
Thanks, glad you find it useful. Please keep watching
I like your videos and I like that you respond to questions in the comments that many viewers would to read an answer to.
One of the things that I am not good at is building a network to get a job in data analytics. I would love of my linkedin page is enough. Thank you for your efforts.
Networking is really hard, and I would also find it difficult. But what you have to do is not think of these interactions as work or getting something form them. If you meet someone then just share details and make an effort to stay in touch. It may lead nowhere or it may lead somewhere or they may connect you with someone else
I'd recommend you join lunchclub.ai
Good place to get better at Networking
@@Shorful Thank you so much!
This is my personal invite
lunchclub.com/?invite_code=shorful
Spot on. Marketing and sales is one of the biggest challenges facing technically strong entrepreneurs in the data and analytics domain.
Thanks for your comment, it's an underestimated area by most technical entrepreneurs
Thank you 🙏🏽 really informative
Glad you found it informative
Excellent coherent video on this topic.
Thank you
Please keep watching
Excellent info. Thanks for sharing
No problem, do let me know if there is anything specific you want to know and i can do a video on that topic
Mate, the UK would be a wonderful market as a data analyst.
keep up the great work
Thanks you, thanks for watching
Please share with friends and family
Great Advice and thank you for placing the indicators for how we should put the ground work in.
Just a quick question sir, how do we market ourselves with 2 years of experience under our belt?
2 years is not a lot of experience if you want to market yourself as a wider data offering, but it is a lot of experience if you are an expert in a tool.
I have helped many people who have limited experience, but because they have been using a tool like SQL or Google Analytics or Fivetran, for about a year, they marketed themselves as experts in this area.
This did limit who they could market themselves to, but they narrowed down on their expertise and found clients who needed experts. From those clients, they built out wider experience and then did more than what they originally.
But, being known as an expert in a tool does mean that you get lots of referrals from clients.
Great video!!
Thanks
awesome video bro
Thank you
Thank you for your video it's really helpful .
I have question, if want to analyze data for client. Should go to their main database and look for the tables i need and linked them and do the analysis. Or as example for SAP i just can use tcode to extract the data to an excel sheet or to access db and do the analysis. Because sometime the client want the dashboard is automated with the database which difficult to get the access from IT team. What is the common request from clients?
That's a common problem.
If clients want an automated dashboard, you either need to connect to the database and read the data for your dashboard, so the client gets updated reports. Or you can get extracts on a regular basis to upload into the dashboard.
The former is the preferred option. But due to security or technical issues its always difficult. The latter is ok, but there are risks of extracts failing and also the report is only updated since the last extract.
On the latter, ideally you want automated extracts as manual extracts have human in the process as dependency, and so have an extra point of failure.
People always say IT restricts access, but I've found if you can work with them on a solution instead of just proposing one, then you can often get access to the underlying database. Where you can't is because of some security protocol or its a third party solution which doesn't allow database access.
Think about creating data pipelines to ingest data into your reporting solutions using connectors in the reporting solutions marketplace, or build one. Most databases have this option.
Thank you for this wonderful information.
I have thoughts about starting a data agency and you provided a lot of great information for me to ponder.
One question that I want to know is. How are pricing determined?
Pricing depends on how you work with clients.
There are two main categories
1. Time and material. Here you just set a day or hourly rate and charge how many days or hours you worked
2. Fixed or Project pricing. Here you work to a fixed price for the work and also ideally a fixed time frame.
I've been doing end-to-end data analytics for around 6 years working as an analytics consultant and statistician , would love to create a company, but as you said, networking and building clients, no clue!
I'm more interested in streamlining processes utilizing python, command line utilities, and batch. Wondering if a market exists for revamping reporting for companies that use wildly inefficient processes.
There is definitely a market for revamping reporting for companies, however, most of the solutions offered are products and companies don't want to invest in tools.
Also, if their current reporting is working, then unlikely they will want to make it more efficient.
My suggestion is that you find a partner who is good at the sales part. You could hire someone, but getting a partner is better, as they'll have the same passion as you.
A department in companies you can focus on for revamping reporting are finance departments. They will work of lots of Excel sheets and usually lots of manual manipulation of the data. So if you can help those teams without disrupting their processes then you could find a niche.
Alternatively, offer to just build new reports, then once inside a company you can see if you can revamp any of their old reports.
I once consulted for a multi billion dollar business who used an Excel report I created as a temporary measure for over 10 years. So, companies rarely revamp if it works
@@Shorful I honestly think people underestimate the prevalence of what I call 'primitive reporting' in some of the largest companies in the world. Excel, Access Databases and Macros. The market is still ripe for would be analysts to offer a service.
Totally agree, I work with some very large companies who just want reporting nothing complex, just reports in Excel or Power BI.
We offer reporting as a service for some clients, but the difficulty is when they want ad-hoc analysis.
So we do a fixed monthly fee for a set number of reports send then either a price per ad-hoc analysis, or they pay a retainer for x number of hours of an analyst time.
It's definitely something that is ripe for solving as a Service.
So here's my question: Why would a company hire a data consultant? How would we market the problem if each company has own wants and needs?
The nature of consultancy is to come into a company, understand their problem and based on your expertise either solve the problem or provide recommendations on how to solve the problem. Typically its the latter.
@@Shorful there's a sports franschise in my city who lacks data needs and support. To convince them to bring me on a board and create a division for data, should I make a case study example and send them ideas of how I'd present raw numbers into beautiful data viz + presentations to explain findings? I got a call from an colleague that they're considering some data roles.
You could show them how you could present data beautifully and explain the data. But what companies want to know is how can the data help them make more profit.
So if you can show through case studies or examples how analysing their data they can make more money, convert more customers, drive operational efficiencies, etc, they should be able to see how hiring a data person is of value to them.
@@Shorfulso I have 2 catamans in koh Samui, Thailand.
I will pay you 50% of net profit that you bring in.
I think fair, you invest nothing in hardware...!
Send me more details at shorful@bedatasolutions.com
Thanks a lot for sharing these useful tips with us. In fact I'm a lecturer and many times my colleagues come to me to ask for some help to analyse or interpret their data. So, the idea came up to me to start a business in Data analytics. I'm good at python and statistics as well as some supervised ML. I would like to know what you think about that. How to start, from where to start and how to present my project to an audience. I'm a pure "academics", and newbie to business
Hi
So if you are a pure academic in this space then use that to your strength and build an analytics business just around advising academics and related people with their data analysis and interpretation of data.
I would start it as a side hustle, I have a video on how to start a side hustle. Then if the business picks up, you can either decide to do it full time, or my advice would be to hire PhD or Masters student part time to help with the business.
As your advantage is being academic, you need to maintain that connection as that is where you knowledge comes from, I would maintain being an academic and also run this business in the side.
ruclips.net/video/HqSGoP9w8hc/видео.html
ruclips.net/video/0jWGPUiKugs/видео.html
ruclips.net/video/A7sWz4wQy-4/видео.html
Let me know if these videos help and if you have any further questions
Hi! Great video. I'm curious about finding clients. Since clients often require a significant amount of data for analysis, smaller businesses might not have access to extensive data sources. How would you approach this issue to contact smaller business?
It depends on the client. It's not the volume of data with small clients but the quality of that data.
If they have a PoS system, then they will have some decent transaction data. Obviously if they have a website then you can also get access to their digital analytics and if they do digital marketing that's also another data source.
I would advise that rather than start with what data you have, start with what business question or challenge you want to answer and then decide what data they have to be able to answer the question.
Good video Bro
Thank you, appreciate it
Hey, you mentioned being able to do 80% of the job yourself. In the event one may not have done a lot of the initial steps, e.g. ingestion, ETL, cloud tools etc, how would you suggest learning and/or practicing these skills?
My recommendation would be that you get that experience doing the actual job rather then through training. Only by doing do you truly know what the tasks involve.
If that's not possible limit the scope of what you offer to those things you have experience in. Then pick up those skills through collaboration with those who have those skills.
I want to ask question on your 1st point do you think somebody who doesn’t know the 80% or lets say he know less than 50% of the work he is doing can’t run a successful agency if what could be the possible challenges they will face ??
There is always the option to partner with people who collectively know more than 80% of the work or just hire people know how to do all the tasks.
So you can still be successful if you can win business.
Again, if you think you can sell, then good, but if you can't you can have a partner who can do the selling or you can hire someone.
The biggest challenge is if something goes wrong, or someone leaves, how will you deal with it of you don't know what they were doing.
Dear Bhaia, how do an analyst or a Data Scientist delivers a project ? Does that come only in forms of jupyter notebooks or Power BI / Tableau files ? Your response will help a lot. Thank you.
Hi Farsim
So it all depends what you are trying to achieve.
Outputs range from a document in either PowerPoint or word or pdf which shows what you found to a python or R script the business will deploy and use the outputs.
Typically though data scientist will create a script which is used by the business to analyse data and get an output, something like a customer score. The data scientist may also run the script themselves if the business wants to do this manually, and then produce the file which contains the customer id and their score.
The output doesn't have to be a customer score, it might be something else such as labelling of images or text.
Other things they may produce would be reports or dashboards which they hand over to the business to use on a regular basis.
If there is anything you want to know specifically, let me know.
Thank you very much. ❤️ This really helps.
Hi DR,
I am from India . I have just completed my graduation . Planning to a start-up of Data Analysis co.
Here I have a question that from where can we find the client those are seeking for the Analysis services ...
Also I am start-up what is you r suggestions how we can start our operations from the start of company to complete of the client service ...
Hi
So I have a video which talks about how to get clients, ruclips.net/video/A7sWz4wQy-4/видео.html
Watch that and let me know if you have any follow up questions.
In terms of how to start your operations, I also have another video on this, ruclips.net/video/0jWGPUiKugs/видео.html. But I see that is the video you have commented on.
Again have a watch and if you have any follow up questions, just let me know.
I am happy to have a discussion about this if you need further help, so connect with me on LinkedIn
www.linkedin.com/in/shorful
with data security being an issue, is it safe to put their company's data on my laptop and use it on open source/free data analytics tools? (like Microsoft azure and Tableau)
So there are 2 separate issues here.
1. If you are going to download any company data to your laptop, absolutely make sure that you do not download any PII data. If you need PII to identify a customer, then on the company's server to create a customer ID from the PII data. But, downloading PII data should be avoided if possible.
2. Security, that is the part that you are referring to. It shouldn't matter what data you are downloading or even generally working on any laptop for a company writing documents, your laptop should have these basic security.
a. Physical security is important. Your laptop should be used and stored somewhere where people can not easily access it, without some sort of access, such as key, card, etc. So if you are using at home, then anyone can't just access your home without a key, so that's Physical security.
b. Software access. So you must have a password to access your laptop, so not anyone can access your laptop. Make sure passwords are not easy to guess. Secondly, the applications you use to analyse the data should also have theiir own login, so PowerBI and Tableau. But at a minimum your harddrive should be encrypted. Finally, make sure you have good antivirus and firewall software.
So, it is safe to have company data on your laptop as long as you are cybersecurity aware, and stay safe online.
@@Shorful that makes sense! thanks!
Thanks sir for such a great video,
I have question
I am planning to learn new skills, and in the options i am having data analytics and full stack developer ( specifically UI/UX designer) . Then my plan is to do a job for some years to gain some field experience and then start my own Buisness,
And here i am unable to take a decision i like data analytics but i am not able to understand how i can start a business in this perticular field how i will bring the clients, cause there should be enough data to analyse and this is the point where i start thinking about going for fullstack developer.
Can you please help me to understand how the buisness model works in data analytics field it would be a great help .
Thanks
My advice is to do a job as a data analyst at a consultancy or agency to understand what services can be sold.
If you stick purely to data analytics, one one end is reporting and on the other end is building propensity models.
Hi, what are your thoughts on providing analytics service to a client as a subscription model ?
So this is traditionally referred to as a retainer model.
The client agrees to set of deliverables or more common a set of resources for an agreed monthly price over a 12 month or longer period.
It's common among consultancies and agencies to offer this to clients.
nice
Thanks
Can you review the coursera Google certificate or tell us what you think about it?
When I get some time, I will check out the syllabus and do a video on what I think of it.
I have successfully developed an end to end process for a data product that i am currently trialling with a few potential clients. However, as somebody who is transitioning from an analyst in FTE to running my own business, i'm unsure how to price my services. Any advice?
Hi
When pricing products vs services it is very different.
For a product, you just have to look at similar products (either related or even those not related) and see how they are priced.
Or if you believe your product is unique, then you can set a price and test it with clients to see what they say.
My only advice is not go in too cheap, as then it's harder to raise prices later, and costs will only increase until you hit scale, at which time unit costs will drop and you can become profitable.
@@Shorful thanks. having read that distinction, mine is more of a service than a product. Though i am trialling it with businesses that do not typically utilise data to reduce operational costs etc. Hence i am struggling to find a similar service that i can reference. I would need to find a suitable price that would not be too low (to avoid the issue of not being able to raise it later very easily) whilst also not being so high where it mitigates the profitability that would incentivise businesses to buy in.
Typically services are just priced as a multiple of a day rate you want, so you just take into account expenses and taxes and then create a day rate based on that value. If you are running it as a company and not a as a freelancer and expect to hire people, then you might want to factor in a profit margin.
But if you want to sell the services as fixed prices, like a product, then think of a competitive price, which covers your base price - which in your case will be people cost - then calculate how many you need to sell to make a profit and then decide if thats doable.
I have a question sir companies already have data analyst business analyst working for them why would they hire a consultancy
There is a whole host of reasons, some listed below, but I am sure there are many more
1. They don't have the skills with their current team
2. They don't have capacity
3. They don't want to do the task
4. The sponsor of the project doesn't want the in-house team to do it
5. The sponsor of the project doesn't know about the in-house team
6. The CEO or someone senior wants to use an external team for whatever reason
7. The consultancy has a reputation, and the company wants to test them
8. The company is thinking of outsourcing some or all of the analytical activities
Like I said, there may be more, but just some I have experienced or know about.
hey brother ,which data sciemce services are mostly needed in companies
Well in my experience, it's mainly predictive models
My dream is I will create a startup but i don't now
Please could you explain in detail
Thank you much
Ok, when I get some time I will create a video that explains the detail of setting up a business
Obviously, as I am in the UK I will speak from that perspective, but most of it should be applicable to anyone anywhere.
If there is anything specific you want to know, let me know in the comments section, and I will address in the video
Please, do a video, I'm really interested
How to select a name for your business ?
I don't think there is a method for this.
One of my companies I named after watching a Marvel movie, another one just from brainstorming.
Choose something that others can say easily and ideally has some connection to what you are doing.
But, I've seen some very successful data analytic consultancies have some random names.
How do you bill your clients ?
The simple answer is I send them an invoice on the amount agreed at the start of the project.
Is that what you meant?
Hi Dr, can I run a company when am not yet qualified?
I am a student and I have completed a few analytics projects.
I would also like to connect with you
Hi
I wouldn't advise running an analytics consultancy without getting some practical experience and also working with a person or team who have done this before so you can learn all aspects of delivering a data analytics consultancy. There are some things which you just can not learn fron books. You need to experience a full project lifecycle to fully understand all aspects.
However, saying that, it's not impossible, you have some options
1. Learn on the job, it may mean that some client projects get delayed as you are learning, but if you have understanding clients, then it should be ok
2. Find a bsuiness partner with more experience or hire someone with more experience
3. Just focus your offering to something you are really comfortable with and only pitch for that work
If this is something you want to do, then go for it, for some people running their own company is natural and they enjoy it more than working for someone else, and the rest you can learn.
To connect with me you can find me on LinkedIn, please connect there www.linkedin.com/in/shorful/
All the best
How can we get client project for our data analytics company??
I'll do a video on this soon.
But you need to
1. Leverage your network
2. Do outbound marketing and sales
3. Utilise partnership programmes.
Kindly make a video on tools or subscription required for this as well. How much money is too much for initial vm setup
Do you mean tools or subscriptions for starting an analytics company.
That would be a good video, thanks for the idea
is there a way to connect with you ?
You can find me on LinkedIn
www.linkedin.com/in/shorful
Interesting but further details needed.
Hi
What other questions do you have, I'd be happy to answer in another video
How much can you charge as an Agency?
The simple answer is that for every person you hire you need to charge 3 x their salary
1 x to cover their salary
1 x to cover expenses
1 x for profit
There is a more complex and sophisticated way to calculate based in different roles and seniority, and the 3x usually applies for agencies which are a decent size with support staff who are not billable.
I am into power bi
Can we please connect?
Sure, you can find me on LinkedIn
www.linkedin.com/in/shorful/
@@Shorful already sent you