Big Data For Small Business


Big Data made Small

Recently I did something very unusual. I posted a review of a lousy meal at a restaurant. It (the meal) was so bad I signed up to TripAdvisor and posted a less-than-positive review. I have no idea if the restaurant owners have ever read my review. If they have, I have no idea if it’s possible for them to send me a message to offer some way of redeeming themselves and winning me over. So far nobody has said the review was helpful, I would at least expect the owners to say it’s helpful. They don’t need to say they enjoyed reading it, but surely it could be helpful? I explained why the meal was horrible, I ranked the various entities fairly, useful feedback.

So this raises the questions:

  • Should the restaurant owners be concerned with reviews, good or bad?
  • Could they do something to get me back on their side?
  • Is there some way they could use the reviews to improve their product?
  • Do people actually read the reviews?
  • For that matter, do the restaurant owners read the reviews?
  • And the final and most important question: does the review data from TripAdvisor get crunched and published somewhere easily accessible? That is, does anyone take this “big data” and really make sense of it and make it easily available?

Actually, I only know the answers to the first three questions. I can categorically say I’ll never set foot in that restaurant. I’ve already told several people how bad it is. My opinion has stopped at least six people from trying it out. However, if the owner was to invite me personally to make up for the terrible first impression I’d probably give it another go. Those six people would be interested in the second opinion.

There is no way anyone would not consider my review to be a contribution to big data. There are a number of ways that this data can be used. It could even be used in isolation (yes, I know big data has Volume/Velocity/Variation but sometimes a single entity is useful). It could be used with other reviews to determine popularity ratings in the area. It could be used to identify weak points in the specific restaurant. It could be used to determine what sort of food I like and the kind of places where I like to eat. Combined with the positive reviews I have subsequently left for other restaurants it should be clear that I like Italian food. More difficult to extract from the data but very important to win me over, I don’t like a noisy stuffed place. I often dine with my wife and we like to talk.

So having described a scenario where a small business could be using big data, I decided to have a good look around the web to see what others were saying. You only have to type in “big data and sm” and Google suggests “big data and small business”. Accept that suggestion and you have a mere 553,000,000 entries to explore. I only read some of them.

I was disappointed.

The majority of examples of how big data can help small business are:

  • Not really big data (come on guys, we can’t simply re-badge relational databases as big data even if they are pretty big). I’m aware that your big data could be my relational data, I’m also aware that all data eventually comes together in a mixture of relational and “other” data. In fact, I’m surprised it took so long to combine Hadoop with our major relational engines. It’s just that when you’re writing an article about big data and you bring in what anyone would call standard relational stuff (e.g. shipping info) and re-badge it as big data I’m not sure I trust your opinions. The people who want to know how big data will benefit their business are, I think, talking about “real” big data.
  • Not really a small business. The problem here is that each country has it’s own idea of what a small business is (see If we go by Microsoft’s Small Business Server that figure would be 50 employees. I think most people would be thinking along the lines of a small family business, a partnership of accountants or lawyers, a restaurant, a store or a small manufacturing concern. It’s not really a question of number of employees or turnover, it’s a pretty subjective thing.
  • Not even data. I agree that this could be a grey area in that just about everything can eventually become data. Example:

Roofing contractors are not the first professionals who spring to mind when you think cutting-edge technology. But some roofers are using big data to drastically cut their costs. A contractor used to take a call, drive to look at a roof and, often, realize it was not a job he could take on. That meant time and money wasted. Smart roofing companies use big data to avoid that expense. A contractor gets a call, takes an address and inspects the roof at Google Earth. If he does the job, he can use Google Earth to check out the roofs of other homes in the neighborhood and offer the owners a deal.

Surely we’re not saying that using Google Earth is Big Data? I would say the fact that he used Google Earth is in itself big data, but all he really did was use technology to avoid a few miles and to save a few minutes.

  • Explanations of how Big Data is for Small Business but without conclusive proof. This is a pretty subjective statement and I apologize if it causes me join the very ranks of those I’m writing adversely about. I simple came away from reading a number of articles with the feeling that they started off saying they would show how apples are apples and ended up showing that bananas aren’t pears. As I said, very subjective.

It seems there is almost desperation to “prove” and “convince” people that big data is for small business, but finding such proof is not easy. I’m not saying that everything is invalid. I’m just saying that 85.6% is invalid.

The problem is, I think I described a pretty good use of big data for a small restaurant at the start of this article, so surely there are other uses.

So … if we were to do a little brainstorming, would we figure out some handy ways for a small business to use Big Data?

Some ways big data can or could be used:

  • The first and most obvious use of big data is to increase revenue if at all possible. The idea is to:
    1. Increase number of customers
    2. Increase amount purchased by each customer
    3. Keep customer retention high, deal with customer dissatisfaction to win them over
    4. Price correctly

All of the above can be assisted by using Big Data, depending on the business.

Of course, different businesses are affected to differing degrees by click data (e.g. a restaurant). In such cases it would be advantageous to tap into Facebooklikes, TripAdvisor reviews etc.

With some businesses advertising can be made more effective and targeted using big data. . Such businesses often deal with large numbers of customers (relative to the business of course). Again, to use my story, do send me adverts of a new Italian restaurant in my area. If my favorite restaurant sent me a special offer it’s all I need as an excuse to head on over. Here’s another hint: I’ve never written a review of an Indian restaurant, so would you target me with an advert of Indian food?

Virtually all businesses can benefit from keeping on top of adverse reviews. Years ago we wrote a review of a logistics company after a failed delivery. The thread is still alive and well and to date the company concerned has never interacted with anyone participating on the thread.

Whole books could be written about correct pricing. If your products are deemed good value for money or not, this could be ascertained from big data.

  • By understanding general things like human behavior and decision processes. By leveraging the output from Big Business Big Data. This will affect certain decision making, such as how to target advertising.
  • Using Big Data to learn how others use Big Data
  • To help you choose service providers. For example, FedEx has a cool product called SenseAware that lets you put sensors on perishable packages. It provides a shipment’s exact location, precise package temperature and information on whether a package has been opened or exposed to light or whether it was dropped, based on the recorded G forces. As more of this sort of data becomes available, comparisons between service providers becomes more analytical.
  • By hiring the right people. This is more of a futuristic thing, and there will be enormous challenges around how the data is actually collected and how unbiased it will be. As a small organization we’re faced with the fact that the wrong employee in the lower ranks can cause havoc. With larger companies there is some safety in numbers. As this process becomes more analytical hiring will be done more as a science than an art.
  • By selling Big Data services

The Challenges:

  1. Where do small businesses get access to relevant big data?
  2. Can they afford to buy it?
  3. How do they go about crunching it? What infrastructure do they need?
  4. If the data is already crunched, can they trust it?
  5. What influence do specific companies have on the way the data is crunched and made available, thereby introducing bias?
  6. If the small business crunches it’s own data, what ecosystem is required to handle things like Data Integrity, refreshes etc?
  7. What if their shoe-string budget introduces error in the way the data is interpreted (this could be valid for large organizations too)?

Oh, and by the way, that 85.6% I mentioned earlier was plucked from the air.

And for the record, I’m still waiting for that restaurant to reach out to me.

White PaperNigel Ivy