Small Business Labs

It's Time to Upgrade U.S. Economic and Labor Statistics

Sometimes we post articles on topics few people are interested in. 

This is such an article.  

But you should be interested because the economic and labor data collected by government agencies no longer provide the information we need to properly understand and manage the U.S. economy.

Over the past week several sources explained the problems with government data collection efforts and made suggestions on how to improve them.

Before we go into them, it's important to note both sources state the biggest problem is we've been cutting the budgets of our statistical agencies for years. 

This has hamstrung their efforts to develop new and better approaches to collecting and disseminating data.

Infrastructure: Upgrading the US labor statistics system is a podcast by Harvard's Future of Work program.  It features former U.S. Bureau of Labor Statistics Erica Groshen, who explains how and why we should improve our economic data gathering capabilities.  

Federal Statistical Needs for a National Advanced Industry and Technology Strategy is a report from the D.C. innovation think tank ITIF.  The key takeaways from this report are:

• The federal government has never felt the need to develop strategic economic intelligence to fully understand the competitive position of its traded sectors or to help support overall economic productivity.
• Years of budget constraints have left U.S. statistical agencies with insufficient resources to effectively measure key elements of the economy.
• There are three key areas with major statistical gaps: subnational data, international data, and data to better understand firm behavior at the enterprise and establishment level.
• As part of a national infrastructure package, Congress should make a one-time appropriation to fully modernize and expand federal statistical agency IT systems. It should then increase annual funding on an ongoing basis.

The lack of timely and complete data on the economy has become a problem for the U.S. 

We regularly see this in our work.  For example, there's still no solid government data on independent workers and most government data on small businesses are too old by the time it's released to be of use.

The economy moves too quickly and changes too often for our old data collection methods.  It's time for the U.S. to make the investment required to collect timely and relevant economic data.

Using Big Data and Machine Learning to Improve Small Business Lending

Intuit recently announced QuickBooks Capital, a new small business lending product that provides users of QuickBooks access to small business loans up to about $35,000.

This new service's lending process is done algorithmically and from within QuickBooks itself. Thanks to big data and machine learning techniques, most borrowers will know whether or not they are approved for a loan in just a a few minutes. 

Lack of credit is consistently one of the top challenges for small businesses. And for new businesses (those that have been in business less than 5 years), the challenges are even greater.

According to the Federal Reserve of New York 2016 Credit Survey: Report on Startup Firms, 70% of young businesses say they need funding to grow. But only 23% of these firms are successful at getting all the funding they are looking for.

    
The main reason these firms struggle is insufficient credit history.

Simply put, they haven't been around long enough to establish a strong enough track record for lenders to be comfortable providing them credit.

And even if they get the credit they're looking for, small business satisfaction with the lending process is not good

Only 48% of successful borrowers report being satisfied with their experience with small banks, 31% with big banks and 23% with online lenders. 

Intuit hopes to improve on all these metrics. 

The QuickBooks Connect credit model is, not surprisingly, based on QuickBooks Online data.

Intuit has over 2 million small businesses that have agreed to allow Intuit to use their anonymized QuickBooks data in an aggregated format to develop products like QuickBooks Capital.

This add ups to over 28 billion data points Intuit can use in its credit model.

(An important side point is under Intuit's user agreement, QuickBooks users own their data and have to expressly provide Intuit permission to use their data in this manner).

This database of income statement, balance sheet, cash flow and transactions data allows Intuit to fully analyze the current financial state of a small business and predict its ability to pay back a loan.

It also means Intuit's credit model has enough data to allow them to lend to young small businesses - even those that have been around less than one year. 

Based on the service's beta customers, this approach is broadening credit availability and making it easier and quicker to get a loan.

According to an analysis by Intuit of its early borrowers:

• 46 percent of QuickBooks Capital customers had never applied for a loan before.
• 60 percent of QuickBooks Capital customers would likely not get a loan elsewhere.
• 90 percent of QuickBooks Capital customers say the loan helped their business grow.
• Almost all borrowers reported being satisfied with the process.

Everyone likes to point out that small businesses are "the engine that drives America's economy".

While this is a bit of an overstatement, it is true they are important drivers of our economy.

Making capital more available and easier to access - especially for young small businesses - will clearly increase economic growth.

QuickBooks Capital is also a good example of how small businesses are going to take advantage of big data, AI and machine learning.

In most cases, small businesses will leverage 3rd party products and services to utilize these advanced techniques. 

Intuit is an Emergent Research (that's us) client. 

Merging Machines and Analytics with the Internet of Things

GE, the large industrial conglomerate, is re-positioning itself as a digital company.

Their recent ads describe GE as an "indugital" and "digidustrial" company. And yes, I agree these marketing efforts are still a work in progress.

They've coined the term "The industrial Internet" to describe their efforts and market. This is basically a new name for The Internet of Things (IoT), a term coined in the late 1990's and defined by Wikipedia as:

... the network of physical objects or "things" embedded with electronics, software, sensors and network connectivity, which enables these objects to collect and exchange data. 

GE sees huge opportunities in this space, particularly by "merging machines and analytics".

According to their Industrial Internet Insights report spending on the Industrial Internet is forecast to reach $500 billion by 2020 and they say the Industrial Internet could add up to $15 trillion of global GDP by 2030.

They're not the only ones thinking this way. According to CB Insights funding of IoT startups has more than doubled over the past 5 years -  from $768M in 2010 to over $1.9B in 2014.

We agree The Industrial Internet/The Internet of Things is a huge opportunity. We even (disclosure) own GE stock as well as stock in several other IoT private and public companies.

We also think the Internet of Things will create new opportunities for small businesses.

We recently partnered with Intuit on a forecast of cloud application use by small businesses. The Intuit chart below shows the average number of cloud applications used by SMBs growing from 3.1 in 2015 to 8.2 in 2020.

One of the drivers of this growth is the Internet of Things.

In particular, we think small businesses will benefit greatly from the IoT by adding applications specifically designed to help them collect, analyze and use operational data to improve their productivity and efficiency. 

We've posted in the past about the Internet of Things finally starting to reach its long forecasted potential. We continue to believe this is correct.

Using Digital Data For Economic Analysis

One of the biggest shifts going on in economics is the growing use of digital data for research.

The Internet and cloud computing platforms are generating large amounts  of digital data on pretty much all areas of business and economics.

In fact, the amount data is growing much faster than our ability to analyze and use it.

In our work we've seen a clear shift to using digital data and away from our traditional sources - surveys and government produced data.

Yes, we still use traditional data sources more than digital data. But we think in a few years our use of digital data will dwarf our use of surveys and government data.

A good example of this shift comes from LinkedIn. They recently held an "Economic Graph Challenge" where researchers submitted research proposals that would use LinkedIn's vast array of data to study various economic issues. 

They selected 11 winning teams and each will get access to LinkedIn data and a $25,000 research grant. The research teams - mostly academics - will crunch LinkedIn data in hopes of drawing some new insights on the global economy and job markets.

The shift to digital data for economic research is very exciting. We have several projects underway on the on-demand economy that are tapping into online digital data. We'll have more to say on this topic in the coming months.

LinkedIn's presentation below provides more information on the winning proposals.

Winners of the LinkedIn Economic Graph Challenge from LinkedIn

It's Not Just Cloud Price Wars Driving Usage

The cloud storage price wars being waged by Google, Amazon, Microsoft have been getting a lot of press this summer.  

As the chart below from TechCrunch's Cloud Storage is Eating the World Alive shows, since March of this year the cost per gigabyte of cloud storage has fallen by more than 50%.

As the article points out, the current price is not far from $0.  

But the price declines are not just limited to storage. Amazon, for example, recently announced price cuts ranging from 28% and 51% across a variety of their Amazon Web Services (AWS) cloud product line.

Cloud application vendors are also joining in.  Splunk, a cloud based data analytics service, recently cut their prices by 33%. 

And Google is telling everyone to expect price declines to continue.  

These cost reductions are leading to more small businesses (and consumers, and big businesses) migrating to the cloud. But it's not just price reductions driving this trend:

• The shift to mobile computing is increasingly requiring cloud storage so files are easily available across multiple devices and geographies.
• Growing numbers of applications are cloud based due to the use of large datasets or the need for computing horse power. Use of these cloud apps are exposing more businesses to the cloud.
• increasing collaboration across teams, geographies and organizations is leading to the growing use of cloud based collaboration tools.
• Cloud uptime and reliability are improving, reducing concerns about the use of cloud applications.

These factors largely explain why we believe the cloud is a transformation technology for small businesses. We explain this in the Intuit presentation Small Business Success in the Cloud.

Labormetrics, Autonomy and Employee Tracking

The Wall Street Journal's Memo to Mobile Workers: The Boss is Watching covers the growing use of labormetrics systems to monitor and track their employees.  Key quote:

Thanks to mobile devices and inexpensive monitoring software, managers can now know where workers are, eavesdrop on their phone calls, tell if a truck driver is wearing his seat belt and intervene if he is tailgating.

The growing use of labormetrics - a buzz word that's a play on sabermetrics, baseball's term for the use of statistics and empirical analysis - is interesting to us on two levels.

First, it's clear labormetrics systems are going to be used in most businesses and across most tasks. 

But it's also interesting as a counter trend to people looking for more autonomy, flexibility and freedom in their work. These are, of course, key reasons people become self-employed.

But it's also increasingly showing up in research on why people like their jobs. The very popular Results Only Work Environment movement, for example, is based on providing employees flexibilty and autonomy.  

The lack of freedom, autonomy and flexibility is also a key reason people don't like their work.

So which path is going to win? Results oriented work environments with lots of freedom or flexibility - or workplaces where labormetrics systems monitor and measure all aspects of work?

The answer is likely both. Those working in high end knowledge based jobs will likely see an increase in the freedom and flexibility in the coming years. Labor competition for people capable of doing this type of work will result in great freedom for these workers.

But expect growing use of labormetrics in lower skilled jobs where labor competition will continue to favor the employer.

The 2013 CRM Market Leaders

CRM Magazine released their annual list of market leaders in the customer relationship management (CRM) space.  They have three award categories based on the firm size of CRM system users - large enterprises, mid-market firms and small businesses.  

Salesforce.com was this year's winner in the small business space, beating out last year's winner Zoho.

Salesforce won the strength of their product direction and reputation for customer satisfaction. Salesforce also has the most complete set of offerings and is very strong for small businesses large enough to have sales teams and sales manager.  

While Zoho didn't win this year, I think they are an excellent choice for smaller small businesses (20 employees or less) and solopreneurs.  In addition to being easy to use and having good functionality, it's cheap if you have more than 3 users and FREE if you have 3 or fewer users.  

Infusionsoft was named this year's rising star. Infusionsoft earned this award in large part due to them earning a very high customer satisfaction score.  

Nimble is also listed as a rising star.  They are leaders in integrating social media data into their CRM system.  If your business is based on building and maintaining strong personal relationships with customers, a social CRM solution like Nimble is well worth looking at.

We continue to be amazed how few small businesses use CRM systems. They are cheap (even free), easy to use and add enormous value by allowing small businesses to collect and analyze customer and prospect information.  

While not all small businesses need them, all small businesses should educate themselves on the space and experiment with using one or two of the products. This doesn't take much time or money and could end up being a business difference maker.

For the 4th year in a row I served as a judge for these awards.  For disclosure purposes, we don't currently have commercial relationships nor have we had a commercial relationship over last 2 year with any of the firms we rated.  

Using Twitter's Promoted Tweets

Promoted Tweets are basically paid ads for Tweets.

They are described by Twitter as "ordinary Tweets purchased by advertisers who want to reach a wider group of users or to spark engagement from their existing followers."

Promoted Tweets have been around for a couple of years, but Twitter limited the number of companies allowed to use them.  They recently changed this policy and now anyone can pay to their Tweets promoted. 

Over the last month Small Business Labs has been experimenting with promoted tweets.  Here are our reactions:

1. Promoted Tweets is extremely easy to use. We were up and running in minutes.

2. Targeting is excellent. You can target your promoted Tweets based on broad interest groups, by keyword, by location and/or by selecting the followers of users.  

We tried a number of different approaches but quickly found selecting the followers of influential users most effective for us. For example, if we were promoting a Tweet about coworking we targeted it a the followers of the thought leaders in the coworking space.  

3. Twitter uses a Google-like auction system for Promoted Tweets.  And like with Google, you only pay for performance, with performance being measure on a cost-per-engagement (CPE) basis. This means you only pay when someone clicks, retweets, replies to, or favorites your Promoted Tweet. You can also choose to pay for followers added as an option.

4.  You can set per day limits, so you aren't exposed to big spend overages.  

For our test we allocated $25 per day for a month and promoted one to three tweets per day. We spent about $350 during the test and we generated a bit over 600 engagements.  We didn't pay for followers.

So we now get to the interesting question - is a Twitter engagement worth the 58 cents we paid per engagement?  

Each business will need to make their own call on this. But for business the answer is subjective and not fully quantifiable. I can easily measure the ROI of clicks that come from my Tweets. But at least for our business, I don't see how to quantify the impact of the other engagement methods - retweets, favoriting and replies. 

Our subjective view is some of our promoted Tweets were well worth the cost, others not so.

Our bottom line is we think we're getting enough value to continue promoting Tweets, but we're going to do so on a very selective basis.

Our advice for small businesses is:

1. If you are an active Twitter user, Promoted Tweets are well worth experimenting with. A test campaign is simple, cheap and takes little time.  

Make sure you measure what you can and minimize your costs. Like with Google Adwords, minimizing cost per engagement is going to be key to a positive ROI. .  

2. If you aren't an active Twitter user, Promoted Tweets are unlikely to be effective and not worth the time and effort associated with a test campaign.

As always, we recommend small businesses experiement with a variety of social media platforms. So if you aren't a Twitter user, consider playing around with it to see if it could be of value to your business.  

What is a Spurious Correlation?

This is a bit more geeky than we like to get here at Smallbizlabs, but the rise of big data is resulting in growing numbers of people being exposed to statistical data and jargon.

A spurious correlation is a statistical term that describes a relationship between two variables that seem to be related (correlated), but happens just by chance or due to an unseen third variable.  

A good example of an unrelated spurious correlation is skirt length theory. This is the belief that stock market trends follow the length of women's skirts. Followers believe when skirts get shorter, the stock market goes up.  When they get longer, it goes down. 

Skirt length theory has been proven to be wrong, but since it works some years (it has a 25% chance of being right in any given year) it continues to have followers.  

An example of a spurious correlation caused by a third variable is the fact that ice cream sales and accidental drownings are highly correlated - they tend to both move up and down in a consistent pattern. 

In this case a 3rd variable - temperature - is the driver. Hotter days result in higher ice cream sales and more people swimming - and, unfortunately - more people drowning.  

Why does this matter?

If you make decisions based on spurious correlations or assume statistical significance where there is none, you will quickly get in trouble. Since there is no causal linkage, the relationship can break down at any point. Going back to skirt length theory, if you bet on the stock market using it you're going to get wiped out.  

It's beyond the scope of this blog to fully cover this topic, but there are two new books well worth reading that do cover it:

Naked Statistics: Stripping the Dread from the Data does an excellent job of describing the basics around statistical causation and the limitations of statistical analysis in layman's terms.

The Signal and the Noise: Why So Many Predictions Fail - but Some Don't is from NY Times rock star statistician Nate Silver. He successfully called the last presidential election and his book covers nicely covers the strengths and weaknesses of statistical analysis.  

In addition to these books, Hubspot's article A Marketer's Guide to Understanding Statistical Significance covers A/B testing, which is widely used (and mis-used) in online marketing.  

Basic statistical skills and understanding is becoming a business requirement. This does not mean you have to be an expert in the field, but if you don't know a mean from a median it's time to learn.  

Confirmation Bias and Decision Making

Regular readers know survivor bias is our favorite bias here at Small Business Labs.  But if we had a 2nd favorite bias, it would be confirmation bias.  This is the tendency for people to favor information that confirms their beliefs or hypotheses.

There's been a flurry of recent articles on confirmation bias:

1.  Doctors, Scientists say fracking critics misrepresent some facts to make their points is from the Washington Post.  It covers the many erroneous claims being made by opponents of natural gas fracking. Key quote:

"Critics of fracking often raise alarms about groundwater pollution, air pollution, and cancer risks, and there are still many uncertainties. But some of the claims have little — or nothing— to back them."

2.  When Bad Theories Happen to Good Scientists from the Wall Street Journal looks at how scientists tend to try to prove their theories rather than see whether or not they correct.  Key quote on how they behave:

"They not only become strongly attached to their own theories; they perpetually look for evidence that supports rather than challenges their theories."

3.  We're all Climate Change Idiots from the New York Times talks about how most Americans have a hard time with issues like climate change. Confirmation bias is a major factor, as illustrated by this quote:

"We also tend to pay attention to information that reinforces what we already believe and dismiss evidence that would require us to change our minds ..."

We run into confirmation bias all the time and struggle with it ourselves.  

Our goal as researchers is to go where the data takes us - and it sometimes takes us to places our clients and audience don't want to go.  When this happens we're amazed the lengths some people will go to ignore or discredit the facts.

Most times confirmation bias just hurts a business, but it can also kill.  Kodak is a great example.  They managed to convince themselves digital photography was not going to go mainstream despite overwhelming evidence to the contrary.

Our advice on dealing with confirmation bias is:

 - try to base decisions on hard data.  This isn't always possible, but the rise of data and analytics means small businesses can be more analytical in their decision making.  Our Big Data category covers this trend in more detail.  

 - get multiple perspectives on key decisions.  We often interview experts from other fields when we're studing a problem.  They bring a different perspective and are more willing to challenge conventional wisdom.  They're also more willing to challenge our assumptions and analysis.  

 - Spend some time identifying and understanding what your views and beliefs are and how they may be impacting your decision making. For example, we're very pro small business and self-employment. Because of this we have to constantly challenge ourselves not to let this bias blind us from seeing negatives in these areas.  

 - Have strong opinions, but hold them weakly.  In our volatile, uncertain, complex and ambiguous world change is the only constant.  Be open to change and accept that even some of our most strongly held business views and beliefs will likely have to change at some point.