Do marketing before building

In the world of web-applications, marketing is often postponed until we have a finished product. I can’t count anymore how many times I’ve been asked to build a new application from scratch. It’s the normal way of thinking and logical. How can you sell something that doesn’t exist yet? Isn’t that unfair?

Consider this: your customer wants you to build a new timesheet application, because he sees a potential market. He has thought out a unique selling point and you like his idea. You know from experience many such projects fail, while your customer may be full of optimism. So what do you do?

While most freelancers would only object mildly and probably end up building the project, I more and more opt for a more extreme approach. I tell the customer he is foolish to pay me X to build this application, because he has no idea if it’s going to work, while we may find out together for a limited budget. I explain that with a simple 1-4 page website you can test almost any business idea in the modern world. There’s even whole crowdfunding platforms, like Kickstarter, who depend on 1 page to pre-sell the product. Only when hitting a set benchmark does the company continue. While many ideas may not work well with Kickstarter, the idea is still solid: do marketing first and only continue upon hitting a specific benchmark.

Most people think me crazy for talking down to my customer and suggesting to not let me build that application. Heck most freelancers would jump at any opportunity to build a web-application, because it will pay them more and bring less hassle then alternate routes. While this may be true short-term, in the long term your customers will respect you more, trust you more and come back with new business ideas to test even if previous ones fail. How many companies and/or freelancers will really help you succeed on that level? Usually my customers know only one.

So how does dry-testing work?

Usually business ideas are mostly in the head of the customer. This will almost be more like consulting or mentoring, but is a vital part to both understand the customer’s wishes and to be able to build a marketing page. Obviously there’s no one right way to do things, but this is a common approach.

1. Work with the customer to define his ideas in words

To start with the idea for a product needs to be explainable in one sentence. Apple did a great job with the ipod “A 1000 songs in your pocket”. Don’t read over this one quickly. Definition is key, especially to dry-test a product. This sentence will often be used as the first thing the potential customer reads and if he or she doesn’t connect with it, or doesn’t understand it, you’re finished.

You may wish to look at tools like the business model canvas or similar tools to help you and the customer further understand all the aspects of his business. For this testing, especially the value proposition and the customer segments are very important.

2. Design and build simple page/website

Depending on your affinity with design, you may even wish to bring in others to help, but remember that most customers don’t really care about design. You can get full web-designs and complete HTML’s for modern and professional looking single-page websites for just a couple of bucks. You can choose to use direct HTML or a CMS platform like WordPress, Kickstarter or any branch-specific platform you can put your page on. Personally I usually prefer direct PHP/HTML as there are better single-page designs to be found in HTML than in WordPress, and all the additional functionality WordPress brings is at this point not necessary. Don’t spent your time on SEO (Search-Engine Optimization).

Remember that your only goal is to get the highest possible conversion rates from visitors to ‘paying’ customers. To measure who pays you can go several routes. It is important to offer the real price from the start. Also try to make the offering simple and to the point. 1 or maximum 2 prices are often sufficient. Making choices too complicated will skew your results.

You can choose to end your funnel at the ‘buy now’ button, but it’s often better to measure real conversions to add a simple form. Adding fake or demo payments is the very best way to ensure your buyers were serious and your conversion represents an actual sale. But let me warn you: actually capturing (saving) payment info like credit card information is illegal in most of the world, so you will probably only capture the e-mail address and name (and even those you can make a checkbox for if the user wants a follow-up or not, or else they might feel even more cheated by you).

3. Add as many opt-in opportunities as possible

The founder of an established e-mail marketing business, who I know well, once told me to find at least 8 ways to convert visitors. Most conversions will be about leaving an e-mail address in one way or another. This may be a newsletter sign-up, free downloadable white-paper or e-book, actual registration or simply to request more information. Now on this simple website it may be a big much to go for 8 ways to convert, but try to go for as many as possible. Any conversion will tell you something about the interest of the customer and will allow for follow-up by e-mail allowing for eventual sales.

4. Measure conversions and results

A very important next step is to make sure you measure what’s happening. Use free tools like Google Analytics or Hotjar. If you chose a single-page website, make sure you record if people are scrolling down and what areas they look at or ‘breakpoints’ they pass. Also very important is to track each of the conversions from 3 separately and get accurate data as to whether or not people are interested.

5. Get visitors

Now your job is to get a representable amount of customers. I suggest to go for at least 1000, as it is generally used as the minimal number to get accurate statistics. Your customer should eventually make the decisions here, but you can help by suggesting advertising platforms like Google Adwords, Facebook or LinkedIn. I suggest not going viral on your own virtual networks at this point, as the best test comes from random visitors, not from friends and family. Obviously there’s many ways to do marketing/advertising and I won’t go into details here, as generally getting a 1000+ visitors isn’t all that hard and just costs a few hundred bucks.

6. Evaluate and determine next actions

You often need a maximum of one week to get the first 1000 visitors and then you can evaluate the results together with your customer. The conversion rates are important here, not the total amount of visitors. Conversions to actual payment (as mentioned in 2) is the most important, but, together with the customer, you can make guesses as to what amount of conversions to things like more information would possibly have. For example you could count 20% of newsletter sign-ups as eventual sales and count them as a sale.

Now you have your results and your conversions the big question is, will this be a viable business? You may not need to help your customer here, but obviously this is all about the numbers. What’s the profit per sale? How does that compare to advertising costs made?

Example case

One of my customers is building a platform to find and book Yoga courses. Now I was too late to go through these steps with her, as I was only brought in when it was almost finished, but this is what I would have advised her to do:

1. Definition

She had already defined her catch sentence: Find local wellness retreats and workshops. I would also have argued to choose to focus on only retreats or only workshops at this point and have her define what “local” is further, by using a specific region for initial testing. Also important to know is where her revenue comes from? In this case from each booking she receives a small percentage.

2. Dry-test website

Since she gets her money from bookings, we’ll focus on that. I would have build a simple website with 4 small pages (easy tracking how far visitors get in the funnel). 1. Homepage with simple search (dropdown on 1 region, ex. Wales) and perhaps start/end date, all captured in url. Homepage  2. search result pages that are fixed. I would probably have opted to put 10 search results (best would be if my customer gets 1-10 real offers, but we can make do with dummy ones if need be) and make it appear as if there are more and capture clicks on more but just show something like ‘page not found’ for now. 3. Book now button will take the user to a simple registration form. To be able to see where the user clicks on I would once again put the name of the dummy retreat/workshop in the url and use it in the payment form on this page to remember the user what he’s paying for. The price may be the same. And the offerings would have different prices on purpose to test. The payment form includes personal information (captured) and payment information (not captured) and upon clicking button [Make Payment] the user would be shown: 4. a page where we apologize for the inconvenience but cannot complete the booking.

3. Mailchimp forms everywhere

To do proper testing we would add button on the homepage (to a Mailchimp standard form) for opt-in to a newsletter. Also we would put in the footer (on all pages) the links “Contact us” (another mailchimp form), “Your retreat or workshop here” (mailchimp). On the search results page we can ask a button for “More information” and use mailchimp again. On the final page we could add a form asking whether or not the user wants information as soon as the booking is available.

4. Google Analytics

For simplicity I install Google Analytics and set-up a funnel for each conversion, where some of the conversions can only be measured by looking at actual sign-ups in mailchimp, but that little manual work is no trouble for now.

5. Specific online marketing

I’d advice to use google adwords for a specific term like “Yoga retreat Wales” and similar terms to get a fairly low PPC (pay per click), while getting the most interested users. We let this run for a week and make the budget high enough to capture at least a 1000 users.

6. Evaluate the results.

Let’s say we get 1 sale, 3 newsletter sign-ups and 2 requests for more information. Checking the newsletter sign-ups and more information requests have 1 duplicate we have 4 requests and will treat this as 1 more sale. 2 sales means a 0.2% conversion. Not that good. Let’s say she earns a 5% share of two $300 retreats, this hypothetical case would have only $30 profit. If Google adwords cost the customer $300, then it’s easy to see that it’s not worth it.

Note that if I cost $75/hour and spent 20 hours on this, allowing her to do things like Mailchimp and Google Adwords set-up herself, she can get this test, with professional help and consulting for only $1800. Building the entire platform, even as a MVP, will cost easily 5-10 times more and will still happen if the results are positive.

By helping your customer through this process they will value you a lot more, more easily recommend you to others, more easily try other business ideas with you if this fails and you can ask a higher price, while still having the possibility of really building the web-application.

And, isn’t it much more fun to work on a project you know will most likely succeed? Not to mention that projects that succeed and grow bigger will need additional work, once again thinking first of you.