It´s all about TRUST

Key Differentiators of Directory Publishers vs Web Search Engines

According to the Kelsey Group, the global sales of directories including print, online and local search is estimated to about 30 Billion USD during 2006. We also know that about 36000 sales reps were involved in the process of earning those revenues. There is another company who single-handedly managed to pass 10 billion USD with less than 40 sales reps. it is no secret that Google strategy is to eliminate human interference in the process of sales. There are different views on whether this is the right or wrong approach in the long run. But I suppose that we can all agree that the sales force is a key differentiator. It does not only affect how things are sold, it also has a significant impact on the actual products that can be taken to the market by respective part.

This article is about turning the key differentiators of directory publishers into competitive advantages. The key differentiators are the sales force, the content and the brand of the directory publishers. Let’s start with the our first key asset – the Sales force.
The more we know about our customers, the better our chances to optimize our sales. That’s where the term data enrichment comes into the picture. I have worked with several customers who initiated data enrichment projects in order to improve their search capabilities. It is known to us all that the best point to start improving search is to start looking at our data. But I have also learned that in the end of most of those data enrichment projects, the biggest beneficiary was the sales force.
Most of you who have been involved in sales management probably recognize the goal of having at least one session with every customer per year. In
the majority of cases, that’s normally all we have time for. One contact a year! At the same time we
know that the majority of our customers are small and medium-sized businesses. And the most typical feature of this category is that they change a lot – one day they furnish kitchens, the other day they do bathrooms and the month after that they do home repairs.

Get to know your rabbit

As a mathematician, I like to count things. Generally it is easier to count things that stand still. In a recently performed analysis of 100.000 commercial web sites, we found that 34% of the web sites were modified just during one week. Not only do they not stand still – they change their positions very rapidly – just like rabbits! In a game, where the rule of the game is one shot a year at every rabbit, we better be prepared before taking that shot.
Luckily we are in a thankful position that there are hundreds of thousands of rabbits around us, so even if we would go out and shoot randomly we still manage to get a catch. But we can neither ignore that there are new hunters in town. The best way to be prepared is to know your rabbit. You got to know what it likes, be able to adapt to its rapid turns, and do your best to increase the probability of a hit. More information about the companies does not only improve search and differentiate our brand, but it can also turn those 36.000 sales reps into well-informed advertisement consultants.

Let’s have a look at an example of one piece of data that has a double effect on both search and sales. Information about the URL of an advertiser is a good way of driving traffic to their web site through the organic search offered by most online directories. If we index the content of the website of the advertisers, we can also offer deeper search into customer and product-specific data. The positive side effect that we have noticed is that prospects who have a website convert better than prospects without a web site when it comes to purchase of online ads. In other words, basic knowledge about the existence of a website, can be used in the process of calculating the probability of shooting the rabbit. Among companies with a web site, we have noticed a relationship between the change frequency of the web site and the size of the orders they place with us. When you think about it, this is quite natural because the commercial web sites with a high change frequency usually contain product information with attributes that changes often, such as price and availability. And companies who earn their money on digital trading are also more willing to spend resources on digital advertising.

A gold mine waiting to be explored

The content and structure of a web site can provide us valuable business intelligence to be utilized by the sales force. Commercial sites with many pictures normally contain products. How many pictures are there? Any specific brands? What is the price range of the products and services sold?

Considering the fact that the web site is a vital information source to the business of directory publishers, the level of attention paid to it is surprisingly low. Most databases we have seen contain only a fraction of the total number of URL’s to commercial web sites. And that fraction normally suffers of relatively high error-rate, in some cases up to 20% of the URL’s are outdated, or faulty. URL Correction and URL detection is the name of two techniques to address this problem. Most URL detection solutions are based on using what we already know about an advertiser to learn more about that advertiser. It is easier to guess the right URL for a company if you know the name, address, phone number and the area of business of that company, which is information that is basic to most yellow pages out there, but quite hard to self-generate for a web search engine.

Eliminating the time-killers of our sales force

Knowing the correct web site of an advertiser, we can start addressing one of the top 5 time-killers of our sales forces, namely: faulty contact information. By validating selected attributes in our databases against the customer’s website, we can point out listings with a high probability of containing faulty contact information and pre-process them. If we suspect that a phone number is faulty, it is of course interesting to have alternatives to that phone number.

This can be done using technologies for entity extraction. An entity is defined by a set of rules, that together specify a pattern. Phone numbers, email addresses and opening hours are example of entities that can be identified fairly easy through pattern matching. Other examples are certifications, brands, emails, addresses, prices and key personnel. The ability to search for patterns in vast amounts of information enables new type of queries in local search. Imagine how this can extend the search capabilities. Now I can finally find the nearest certified tattooist with opening hours that suites my agenda!

Dear Customer, which keywords do YOU want to purchase today?

Knowing more about the customers also creates a good foundation for trust between the sales rep and the customer. In my world, questioning a plumber about the keywords he wants to buy or the headings he consider as relevant within our taxonomy simply does not make sense. How would he know? If we content ourselves by just asking him, we will probably get some suggestions from the top of his mind. And that piece of his mind will probably cover only a fraction of his entire business. How do we assure that the knowledge about his business that is necessary for correct exposure of him is transferred to us, during the single shot we have at the rabbit?

The nature of the phenomenon called the long tail is a direct result of the diversity of the consumers. And it is reasonable to assume that this diversity is reflected in the orientation of the suppliers. Most small and medium businesses live in a belief that they are unique, or somehow oriented towards a certain specialization. To connect the right buyers with relevant sellers, we should not only cherish this diversity, but illuminate it to our best abilities. Suggesting the same set of keywords to all companies within a category stands in direct conflict with what we said about cherishing the diversity. The same thing goes for the faulty procedure of suggesting the same related categories to all companies within a certain category. There are new solutions available that can use existing categorization as a starting point and generate representative keywords and relevant headings not only for each category, but even for each company.

The clockworks of the human mind

Human beings are better at validating information they are provided then being involved in creation of new information. It takes less time to have a prospect verify the content of a customized package purposely assembled to cover his specific specializations, than to involve him in the process of creating that package from scratch. By using technologies for automatic keyword suggestion and auto-classification, our sales reps can avoid spending time asking basic questions about the advertiser, trying to create and assemble new information manually in the middle of the rare event of live interaction with the customer. We can use that time providing useful recommendations and valuable information relevant to the core business of the advertiser instead. Why not tell him about what’s hot right now within his specific line of business, or give concrete recommendations about how he can optimize his investment in terms of exposure, and back that up with easy to grasp dash-boards and facts?
Imagine our sales reps being able to state the following magic words in a sales conversation: “Here is what I think makes you unique (in terms of a package of representative terms for your business) paired with information about how people search for this. And by the way, I noticed on your website that you have started selling bathrooms. How about rearranging your headings according to the list I have prepared for you so we can increase your exposure with an additional 20%?”

Looking local by being local

It is hard to give a local impression when lacking knowledge about the basics. Being local implies being well-informed. Together with a well-informed impression comes trust. Together with trust come better conversion rates. As a small or medium sized business, I would rather spend my advertisement budget on someone who seems to understand my business and does the job for me, than someone who’s main proposal is a set of do-it-yourself tools.
Another area where trust seems to gain importance is search. I have had the pleasure of working with customers in Russia recently. And it has been interesting to learn the impact of cultural factors on the way people search for information. If I want a plumber in Sweden, I probably just look in the yellow pages to generate a primary short list of possible vendors. In Russia, the same short list is generated by initiating a series of calls to family and friends. You do not only get your short list that way, but also reviews, price information, bargaining procedures, and lots of other valuable “user-generated” content. It’s a very neat implementation of web 2.0 without actually using the Web!
Do note that the contact information (or other basic listings) about the vendor is of less importance in the generation of the short list, and that the directory search for the number actually occurs after the decision for which vendor to use has already been made!
So if the nature of people is to make their decisions based on trust, why not integrate that behavior in the process of search? It would be neat to be able to search for a plumber that is actually recommended by one of my friends in my contact list, or maybe even a friend to a friend, wouldn’t it?

The future of Local Search

The best way to predict the future is to learn from the past. Thomas Edison is known by the masses as the inventor of the lamp. Edison is known as the great innovator who brought the light into our houses starting 1882. It is however known to most physicists that the technology behind electric arc lamp was discovered 40 years earlier, back in 1845. The genius of Edison was more about combining technical inspiration with commercial perspiration when he paired the arc lamp with a simple but scalable solution for generation of electricity and an easy-to-grasp business model.

I believe that the future of local search goes beyond search as we know it today. Search is not about enabling functions. The true search experience is about enabling perspectives that depends on each users point of view. It’s all about understanding the elements of interaction between people, businesses and places, where the relevancy of the answers is based on analysis of relationships, patterns and contexts. We are only in the beginning of this exciting path. So I would like to end by quoting Mr. Ingvar Kamprad, the founder of IKEA: “Most of my work is yet to be done – oh, what a wonderful future!”

June 2007, Michael Mokhberi