Fix­at­ed on fix­ing: Win­dows & B2B eCommerce

Web­sites are peo­ple too

Your Ecom­merce is much more than the hid­den cogs, bolts, code and pix­els that make it work. It’s a liv­ing arti­fact that com­mu­ni­cates with oth­er humans, on your behalf.

A dis­tinc­tive voice that, done right, can tell your vis­i­tors exact­ly why they should become cus­tomers. Why you are dif­fer­ent, spe­cial, per­fect for them.

Some­times we for­get this, as we’re too busy swim­ming through con­flict­ing pri­or­i­ties, and the voice rep­re­sent­ing us becomes pale, dis­tort­ed, inac­cu­rate, or plain ugly.

In Social Sci­ences, the Bro­ken Win­dow The­o­ry essen­tial­ly looks to explain that when nobody gives a crap, will­ing­ly or not, prob­lems scale quickly.

The land­scape talks to peo­ple. When a bro­ken win­dow is not fixed, the remain­ing ones will be bro­ken soon, sim­ply because it con­veys the idea that no one cares. And that idea is per­va­sive enough to stick and become a reality.

Unfixed prob­lems are con­ta­gious. Both a symp­tom and a disease.

Design­ing for gut feelings

Most of us don’t come up with lists of pros and cons, detailed com­par­isons or deci­sion matri­ces to choose ven­dor one or ven­dor two. We decide based on a ton of fac­tors that are sub­con­scious­ly processed, and then ratio­nal­ize them to con­vince our­selves that ​“good enough” equals ​“opti­mal” (see Bound­ed Ratio­nal­i­ty).

Pro­vid­ing a façade with­out bro­ken win­dows – ide­al­ly one that’s beau­ti­ful and thought-through as well – is key to cre­at­ing a good first impres­sion and remov­ing block­ers that alter the heuris­tics on con­ver­sion deci­sions. Medina’s Brain Rules state that emo­tion­al­ly charged events per­sist much longer in our mem­o­ries and are recalled with greater accu­ra­cy than neu­tral memories.

You don’t nec­es­sar­i­ly have to be the best to be the per­fect choice. ​“Good enough” works. And when it’s deliv­ered frus­tra­tion-free, it cre­ates trust.

Trust is a gut feel­ing more than a ratio­nal process, and visu­al design affects emo­tions in a very pow­er­ful way, per­haps more than any oth­er stimuli.

Jason Putor­ti
For­mer Head of Design at Mint​.com

Mon­ey on the table

Every time you post­pone a fix, you incur a debt. Keep adding fix­es to the ledger, and it becomes very hard to pay. On top of that, it makes you lose mon­ey. There’s plen­ty of research on the subject:

• 88% of users are unlike­ly to return to a web­site after a bad user expe­ri­ence (Spi­r­a­lyt­ics)
• 17% of cart aban­don­ment is due to errors on the web­site (Bay­mard)
• 94% of rejec­tions come from poor design-relat­ed first impres­sion (Paper)

If you read our pre­vi­ous insights, you know we’re big on the 95 – 5 rule that states that in B2B, only 5% of your poten­tial buy­ers are active­ly look­ing to make a pur­chase. Busi­ness­es can­not afford to lose such a small per­cent­age to poor experiences.

Let’s see the glass half full

This is a mas­sive oppor­tu­ni­ty to stand out. More numbers:

• A company’s cus­tomer base can increase by 1.11% for every 100ms decrease in home­page load­ing speed (Hobo Web, 2023)
• Only 18% of cus­tomers feel sat­is­fied with a company’s effort to per­son­al­ize expe­ri­ences (Adobe, 2023). In oth­er words, 82% of com­pa­nies suck at per­son­al­iz­ing their customer’s expe­ri­ences. The bar is very low.
• 72% of cus­tomers will tell six peo­ple about a good expe­ri­ence (Spi­r­a­lyt­ics)

In short, basic, good enough work puts you above your competition.

As an agency we often see this dras­tic sce­nario: Clients are so tired of their plat­forms not deliv­er­ing, their web­sites being slow and hard to update, and their cus­tomers com­plain­ing about bugs, that they decide to throw every­thing away and start over.

Some­times it’s a good idea. But many times, it’s not. You don’t need to imple­ment crazy expen­sive and com­plex redesign, or seri­al­ly adopt new tech­nolo­gies, or have an end­less bud­get. Start with fix­ing what needs to be fixed, and work your way up in small improvements.

This is what a bro­ken win­dow looks like

Think of your own shop­ping expe­ri­ences. You try to shop for some­thing online, and often find dis­cour­ag­ing ele­ments like:

• Slow load­ing times
• Bro­ken or gener­ic images
• Key­word over­loaded prod­uct descrip­tions that don’t real­ly say any­thing to you as a human
• Search box­es where it’s impos­si­ble to find anything
• Poor mobile expe­ri­ence with web­sites that are bad­ly optimized
• Bro­ken links
• Bro­ken fea­tures. How many times do you stum­ble into bro­ken sign up forms, pass­word recov­ery issues, crash­ing captcha, and more?

These things say loud and clear ​“we don’t care”. But we do. We real­ly do. 

So let’s fix them.

Quick things you can do on a budget

Who has two thumbs and loves quick wins?

1: Find the problems

Microsoft Clar­i­ty can help you spot UX behav­ior and track cus­tomer activ­i­ty. Install it, let it run for a few weeks, and see what you can learn from it. It’s excel­lent and it’s free.

Back in the 2000 Nielsen found that test­ing with five users is enough, and it still holds true. Bring five real cus­tomers. Mix brand cham­pi­ons with users that left your web­site frus­trat­ed. Be gen­er­ous and reward their time, and have them run mul­ti­ple small exper­i­ments to see how they use the web­site. Explore the objec­tive results (Did they per­form the task suc­cess­ful­ly?) as well as their emo­tion­al response (Was it easy? Are you pleased with the outcome?).

Repeat this process at least twice a year. Get­ting real feed­back helps leav­ing per­son­al pref­er­ences aside and focus­ing on what actu­al­ly mat­ters to your customers.

Last­ly, have some­one on the team run man­u­al test­ing of every­thing on the site. There are many opin­ions on how, how much, and how often you should do it. Our rule of thumb for test­ing is:

• Low traf­fic web­sites (> 50k month­ly ses­sions): Test all your key con­tent, sec­tions and shop­ping cycle end to end on 5 pop­u­lar browsers (desk­top and mobile), once a month.
• Mid sized traf­fic web­sites (50 – 200k month­ly ses­sions): Text every two weeks and grow that list to at least ten devices. You can use tools like Browser­stack to test remotely.
• If you have more traf­fic, you might want to con­sid­er a full or par­tial resource that can stay on top of test­ing on a dai­ly basis, and in per­ma­nent touch with your devel­op­ment team or your AOR.

Remem­ber we can help too and the price tag is a no brainer.

2: Under­stand the problems

Where does it hurt? Dig deep­er to iden­ti­fy the small issues that are caus­ing fric­tion. This could be any­thing from a slow-load­ing page to a con­fus­ing nav­i­ga­tion menu.

• Cre­ate a KPI dash­board with Look­er, Data­box or similar
• Run light­house per­for­mance tests
• List prob­lems iden­ti­fied with Microsoft Clarity

Not all prob­lems are cre­at­ed equal. Pri­or­i­tize fix­es based on how they impact on cus­tomer expe­ri­ence and busi­ness outcomes.

• Run a quick matrix with two axes: Effort and impact. Effort is more or less objec­tive. Impact depends on how hap­py you can make your cus­tomers, or how exten­sive the prob­lem is.
• Plan a rea­son­able dev cal­en­dar to work through the list.

Focus on mak­ing small, incre­men­tal changes rather than attempt­ing a com­plete over­haul. This approach reduces risk and lets you to see the impact of each change more clearly.

3: Solve the problems

“Go big or go home” doesn’t apply to web­site main­te­nance. Imple­ment small incre­men­tal changes and keep deliv­er­ing nice sur­pris­es to your cus­tomers. Atom­ic changes accu­mu­late and become expo­nen­tial very quickly.

• Con­sid­er Agile or Growth-based sprints. Batch tasks in short groups of two to four weeks. Have a release sched­ule. Make sure you keep an eye on the ana­lyt­ics avail­able to fig­ure out what’s mov­ing the needle. 
• Break­ing down projects into small­er, man­age­able tasks and iter­at­ing con­tin­u­ous­ly lead to steady progress with­out the need for mas­sive over­hauls. Sprints → feed­back → iteration.
• Not every change imple­ment­ed works as expect­ed, and that’s okay. Know­ing this reduces inter­nal team pres­sure, fric­tion and frus­tra­tion, and makes you bet­ter at han­dling change fast.

Last, make sure you let your cus­tomer know about what’s new and bet­ter. Con­vey urgency and thank­ful­ness in sup­port requests and fol­low up once the issues are fixed. Remem­ber that solv­ing a com­plaint in the customer’s favor can lead to a 70% reten­tion rate.

Mend­ing the fence

Beau­ty and func­tion – the two essen­tial pil­lars of gen­er­al pur­pose design – are good for busi­ness because they intrin­si­cal­ly say ​“I care”, gen­er­ate trust and facil­i­tate post-ratio­nal­ized gut-based purchases

Every new unfixed prob­lem incurs a tech­ni­cal debt, and makes you lose sales. It hits twice, and twice as hard.

Fix your win­dows in small chunks and make plans to make things more beau­ti­ful lit­tle by lit­tle. You don’t need a big bud­get. You’ll save mon­ey in the long run, get bet­ter com­mer­cial results, and lov­ing­ly show your cus­tomers that you can – para­phras­ing our favorite bea­t­le – han­dle them with care.