How does Dataro compare to me creating my own campaign lists based on my own segments?

There are a number of advantages to using a machine learning based approach to create fundraising campaign lists compared to typical 'segmentation' methods like RFM analysis. 

• You can think of machine learning as a supercharged version of segmentation: hundreds of donor attributes can be analysed, and then mixed and matched together. All of this extra information is going to give you a more accurate prediction for how donors are going to behave.
• This means you’ll be able to do things like create smaller campaign lists that raise more money, identify people that you otherwise would have missed from your segments and incorporate donors from other parts of your database that you didn't know were likely to give. 
• One example of how machine learning compares to segments is where you consider a typical ‘bad’ segment, such as a cohort of long-lapsed donors. If you were to contact the whole segment you will likely return a poor response rate – that’s because segmentation doesn’t consider sufficient information to reliably tell the difference between the person who might give and the person who won’t. Using machine learning you can better identify the people with a higher probability of giving, which means that you can improve fundraising performance across the board – even for your ‘bad’ segments.
• Machine learning finds the most ‘optimal’ solution to a problem given the data available. Instead of considering a small set of factors (e.g. recency, frequency, value of gifts), machine learning can take into account hundreds of factors, and discover new and novel drivers for cohort selection that you may not have considered or that may even surprise you. It can even open up the possibility of new types of fundraising campaigns, like proactive retention campaigns based on a donors likelihood to churn.