SnipSnap & Wizard of Oz Engineering

I still get asked every week how hard was it to build the machine that powers SnipSnap? How long did it take you to master OCR and image recognition and machine learning, which appear to make the instantaneous, magical scanning of coupons possible.

The truth is: a) I didn’t build squat — that was all Kostas and our dev team. b) We actually didn’t master any of these up front. We instead became masters of what is known to some startups as Wizard of Oz engineering. Peel back the curtain and, voila, you’ve got people pulling the levers.

We realized early on that it would take us months (maybe years) to get any kind of funtional auto-detection and parsing working. So we cobbled together a solution that leaned heavily on Amazon’s mechnical turk. We’ve since gotten better and better at automating some pieces of the process, and have actually developed some innovative tech in the OCR space in the process. But the story of how we got our first prototype and V1 of the app to market is a useful tale of trial and error … and more error, that I hope at least one fellow entrepreneur out there will appreciate. I wrote it up for Fast Company. Here’s the opening excerpt.

When my iOS app SnipSnap was accepted to DreamIt Venturestwo years ago, it was little more than a screencast and a high-fidelity prototype built on Keynote templates. Vaporwear. We were planning to build a fairly sophisticated OCR app for coupons and had zero technology. But we learned you can overpromise like this with early features–if you know your way around Amazon’sMechanical Turk service.

Check out the full SnipSnap Coupon App story on Co.Labs.

Read More

All Aboard the Warby Parker Marketing Train … er, Bus

This weekend I took Caroline and Charlie on their first Patco train ride into Philly. We didn’t have a very clear itinerary in mind. The mission was simply to get out of the house for a few hours, so Ana, who was post-call, could get some shut eye.

After a thrilling, bench-seat-thumping ride into the city, we found ourselves wandering over to the Warby Parker Class Trip bus parked on Rittenhouse. Not exactly the most kid-friendly destination, but my startup fanboy curiosity got the better of me. Why, I’ve been wondering, would an ecommerce giant like Warby Parker need to do a mobile road show?

I’m a huge fan of Warby Parker. On a meta level, I’m inspired by their bold choice to take on eyewear monopolist Luxottica with $95-flat-fee pricing. I’m crazy about their at-home trial program — from the selection of 5 frames online to the sensual unboxing of their silk-box mailers. In terms of UX, I’m not sure any e-commerce retailer comes close. And their virtual try-on service, where you can model a pair of specs using your webcam, is nifty. Oh yeah, the glasses themselves are pretty sharp, too.

But despite all my enthusiasm, and two batches of at-home trials, I couldn’t pull the trigger on a pair. Yes, the frames are stylish, but uh, maybe a bit too much. It took a bricks and mortar store with the full Warby Parker line on display (Art in the Age, in Old City Philadelphia) for me to finally fall in love with one of their frames.

So for this reason alone, I was already pre-sold on having real showrooms for WP. But why throw wheels on it and travel from city to city for the Warby Parker Class Trip? Turns out creating a showroom like this in high-traffic spots like Rittenhouse Square becomes a kind of spectacle. The bus was packed — a good thing, especially when Caroline and Charlie started roughhousing with the specs. They had a photo booth for momentos (and reminders of which frames you should buy). And hey, no rent!

Warby Parker isn’t the only e-retailer pulling a clever online-to-offline-to-online magic trick. Walking back along Chestnut, I passed by another Philly startup that’s crushing it: Kembrel. Not only is Kembrel one of the best currators of flash-sale accessories and clothes, but they made the smart move to locate cheap, unrented storefront space to set up quasi pop-up shops. A great way to get out the word on their business, but also a physical spot to house their inventory. I’m not all that familiar with the in’s and out’s of their business, but I believe they handle fulfillment right out of the Chestnut Street store.

Lately we’ve been taking to the streets to promote SnipSnap offline, too. We started at the Cherry Hill Mall, handing out coupon bundle cards — “$200 in coupons for stores at this mall! Scan this code to import them.”

So far 25% of the people we hand these cards to download the app and import the bundle. That’s our highest conversion rate of any marketing tactic … by far. Of course, our poor intern was also kicked out of the mall one time, so we may need to revise the strategy soon. Kiosks? Giveaways?

Whatever we end up doing, I just hope we can be 1/2 as effective at creating a spectacle (and indoctrinating evangelists) as Warby Parker and its bus.

Read More

10 Things I wish I’d known heading into startup boot camp at DreamIt Ventures

My time at DreamIt Ventures, a Philly-based accelerator program for startups, was beyond rewarding. I feel like I got an MBA in three month’s time — and they paid me to attend classes. This is where I set out to build my startup, SnipSnap (coming soon to an iPhone near you).

It’s hard to boil down such an intense, round-the-clock experience into one blog post, but I’d like to at least try to summarize 10 of the major lessons learned.

10. Get to know your startup neighbors. I learned more from the 14 companies we worked alongside than everything else about the program combined. I now count some of the Fall 2011 CEOs among my most trusted confidants — the guys and girls I send my term sheets to for feedback, or who I’ll solicit advice on how to handle a thorny legal issue. I just wish I’d started befriending more of them sooner. There are still a handful I never got to know, and that’s one of my biggest regrets of the program.

9. Practice your investor pitch and term-sheet negotiations. File this as something we didn’t do much of, but I wish we had. Yes, we practiced the Demo Day pitch ad nauseum, but given that all of the DreamIt partners are seasoned investors, I wish we’d done mock one-on-one investor presentations and TS negotiations before leaving.

8. Take care of all your legal needs before graduating. If a law firm is donating its legal services, make sure to finish as much as possible before the real-world meter kicks in.

7. Go to the evening social events. Trust me, I know how hard this is for parents. And I can say that with one toddler at home and a baby born mid-program, it wasn’t easy for me at all. But this 3-month period is one of the best shots you’ll have to immerse yourself in the city tech scene. Don’t miss it.

6. Begin fundraising on day 1. That’s hard, of course, if you’re still fleshing out just what your product is. But if you hope to have any kind of commitments come graduation, you best start early. Even then, you’ll be lucky if you can say that you have 50% hard-circled by D-day. It’s now three months after Demo Day, and I’m only now beginning to solidify our round.

5. Ask other teams for help. The teams that seemed to grow the most, and engage the entire community, reached out to other companies for feedback. Occasionally they would even enlist an engineer or salesperson to help them solve a problem. This is also partly important because …

4. Everyone in the incubator space is a possible hire. It’s a fact: Most of the accelerator companies will fizzle in under a year. Which means there’s a ton of talent out there looking for a new home. I enlisted some DreamIt grads from past years to help on SnipSnap. In some cases, companies recruited employees from other teams before the program even concluded. While this may sound like it’s crossing a line, I think everyone will agree that if you’re able to poach an employee, they were already looking to be poached.

3. Meet with your team daily. Again, file this under “do as I say, not as I do.” I wish team SnipSnap had met this regularly. The teams that did were operating on a whole other level. My new team is indeed having a daily huddle.

2. Minimum Viable Product ≠ prototype. It’s so easy to get sucked into the Eric Ries, “Lean Startup” mindset, where you strip down your product to the bare essentials, that you start to think it’s ok if a certain feature doesn’t technically work. We made this mistake with a half-dozen parts of SnipSnap. When the time came to actually fix them — that is, get them working — we were so stuck on the MVP mindset that it inordinately painful to get un-stuck.

1. Find the right co-founder(s). This one is more than a little painful. My original co-founders for SnipSnap didn’t stick around long past our DreamIt graduation. They had another startup, which was demanding more and more of their time. It was a good business, and I don’t begrudge them wanting to invest themselves fully in it. And in the end, it worked out best for me as I was able to build a new team that was much more passionate about the app we were building. We quickly found a groove and I’m thrilled to say that this new team is humming like a well oiled machine. In many ways, it’s the one I wish I’d applied to DreamIt with.

If I could do it all over again, this last point is really the one I would drive home. It’s tempting to apply to an accelerator with a team pulled out of thin air — people you know, but haven’t worked with on a startup before — but it’s a dangerous gamble. If possible, work together for a few months in advance. Get to know each other. Even if you don’t start finishing each other’s sentences, at least make sure these are the folks you want to spend the next 5+ years of your life with. A good friend — one of those DreamIt CEOs I mentioned — likens the search for a co-founder to a courtship. Very true. And to that point, whatever you do, please don’t propose co-founder marriage before you’ve finished your first date.

Read More

Resolved: To Shop Local (and price match) in 2012

When Santa failed to bring us naughty kids a new double baby stroller for Christmas this year, Ana quickly set to work on finding a way to get us the preferred model with at least a 20% discount. Not two weeks later, mission accomplished. Only problem: It was at

Normally I don’t think twice about ordering online, but lately I’ve noticed a domino chain of stores closing in my town, Haddonfield. Last week, I quietly resolved to make 2012 the year I actually try to shop local.

It’s hard, of course, given that nearly everything we’re inclined to buy these days is available via iTunes or Amazon, but still: Something like the Bob Duali baby jogger stroller we can easily find in town. And sure enough, the Haddonfield Running Company had exactly the right model — only for about $150 more than the REI price.

Just for the heck of it, I walked in today with a printout from the REI site and asked if they would price match. “We can do that, sure.” No argument at all. They thanked me for giving the local mom-and-pop shop a shot.

Up until now I’ve never tried this anywhere but a know price-matching retailer (e.g. Buy Buy Baby), but after that happy-making experience, I may just have to walk into every ol’ store in town with a printout — if not a SnipSnap screenshot.

Read More

Meet Charlie

Charles Henry Mann

Born: 11/11/11
7 lbs 12 oz

Very proud of mother and son for sticking the numerologically awesome birth date.

Long live my son.

Read More

A Hyperlocal Post-Mortem: Lessons Learned from InJersey

Last week I made the heart-wrenching decision to stop publishing InJersey, a network of hyperlocal websites that I’d founded two years ago. It was particularly difficult because I’d grown to love the community of folks that supported our sites — from the site editors under the employ of Gannett to the community contributors to the far-flung partners and champions of the local-local journalism. To help get through the mourning period, I penned a post for Street Fight (reprinted below), a site that has emerged as the blog of record for the hyperlocal industry. The comments that were posted to both it, as well those on a fantastic article by Mallary Jean Tenore on Poynter (posted more or less simultaneously) and a follow-up Street Fight post, are a testament to how strong and uplifting this community has been. Thanks in part to this kind of camaraderie, and the promise that community-centric journalism still holds, you can rest assured that InJersey won’t be my last foray into the world of hyperlocal.

When we made the decision this week to shutter — a network of hyperlocal sites across the garden state that I helped build, nurture, and raise like a child — my biggest fear was that the effort would be branded a failure. In the age of Twitter, I was braced for the #epicfail hashtag. It came instead via Slate, in the form of a Jack Shafer missive, where he lumped my baby in as yet another vote of no-confidence in a “hyperloco” business model:

“The hyperlocal cemetery—where Bayosphere; the Washington Post’s LoudounExtra; Allbritton’s TBD; Backfence; and the New York Times’ New Jersey experiment, “The Local” are taking the dirt nap—got a new tenant today, as Gannett folded hyperlocal venture InJersey.”

The truth is that my colleagues and I viewed InJersey as a hugely successful experiment. Lest that sound like empty spin, here are the facts: We launched two years ago on a shoestring budget, and grew to 17 town blogs, 2,259 members, and 90 community contributors. Though we were never able to afford a full-time staff, we recruited site editors from the stable of reporters across the Gannett New Jersey newspapers — the Asbury Park Press, Home News Tribune, Courier News, Daily Record, Courier-Post, and The Daily Journal — who built up a tremendous online community. Some town sites, such as Freehold InJersey, earned as many as 65,000 pageviews and 47,000 unique users in a month. We sold sponsorships onto the town sites, and advertisers were thrilled with the response. We were profitable.

Let me repeat that: We were profitable.

But, alas, we never generated enough revenue to pay for a full time reporting (or sales) staff, and given the contractions of the newspapers (including a particularly tough round of layoffs last week), we couldn’t sustain the effort.

Still, as I said in my parting post on InJersey, even though we decided to suspend publishing, we did so knowing full well that the larger hyperlocal movement that we belonged to is as vibrant and innovative as ever. I’d hate for this act of infanticide to be minimized into a footnote on the wikipedia “hyperlocal” page. And so I want to share with you some of the major lessons I learned from InJersey.

1. Build Cheaply. Shortly after launching, Jeff Jarvis asked me to speak about the tools needed to launch a hyperlocal site at his New Business Models for News conference (or as I liked to call it, “HyperCamp”). I told the audience there that anyone can launch a killer community site for a few thousand dollars. Most can and do spend next to nothing. To drive home this point, I compared the budget to build InJersey ($3,720) with the cost of the movie “Paranormal Activity” ($11,000). It can be done.

We used WordPress for our CMS and BuddyPress for messaging, community registration, collaboration, and more. TribLocal uses much the same approach with their network. I have tremendous admiration for the kind of custom CMS that Patch has engineered — I’ve seen it; it’s impressive — but this doesn’t yield much of an advantage, IMHO. As hyperlocal swami Mark Potts says “It’s not about technology, it’s not about journalism, it’s not about whizbang Web 2.0 features. It’s about bringing community members together to share what they know about what’s going on around town.”

2. Recruit Via Twitter. I’m still so thankful for the Twitter DM gods for sending me Colleen Curry. She contacted me out of the blue, back when I was putting the finishing touches on our first WordPress theme, asking if we needed any writers. Luckily, we were able to bring her on to not only cover the blog, but also as a reporter for the Asbury Park Press, and she’s been an animal ever since — covering town hall, school boards, breaking news, dangerous intersections, and even stray turkeys. (Story of her hiring, in her words.) Her Freehold site was our top performing hyperlocal. At the same time we shuttered InJersey, I’m proud to say she was promoted to become USA Today’s New Jersey correspondent (#silverlining!).

3. Partner with Everyone. This may sound obvious, but I’m still surprised at how many hyperlocal sites start off all on their lonesome. We actively reached out to other startups (SeeClickFix, Outside.In, PaperG), local groups (United Way, Rotary, Libraries), and residents (liberals, conservatives, high-schoolers, old timers).

What made our approach to hyperlocal different than others (at least when we launched) was that we offered open registration and the ability for anyone to publish. Yes, we had our fair share of off-the-wall and inflammatory posts, but we took them down, contacted the contributor, and coached them on how to revise. It wasn’t a perfect system, but there was a great deal less friction than the Times’s “The Local” faced by editing every contributor’s post via email. We’d hoped to have 50% of our posts generated by the community, and while we only got half-way there, it wasn’t for lack of trying.

4. Local Advertisers Won’t Self Serve (but they will support local coverage). Like so many digital startups, we’d hoped that tools like PaperG’s Flyerboard (also looked at AdReady and AdTaily) would give advertisers the tools to easily build, schedule, and purchase their ads all by themselves. Yeah, didn’t happen. That said, there was definitely an appetite among local advertisers to support the kind of reporting and coverage we were providing. Still loved using Flyerboard, we just had to sell them, create the ads, and schedule it all ourselves.

5. Maps are eye candy. No more, no less. We kind of fell in love with Patch’s first map treatment on their homepage. So much so that we built a WordPress plugin to ape it. Only when we went live did we realize nobody was clicking on it. Next we tried Outside.In‘s full-page maps and widgets, which were even cooler — they automatically tagged our content with geo-locations based on contextual clues. Amazing magic trick. Alas, nobody clicked on them either.

Original Patch Map

InJersey's original map widget

(I suspect that the value of geo-tagging content and location will grow as we find new ways to integrate it with a mobile experience — e.g. presenting you with news and deals based on where you’re standing at that moment. Now that Outside.In is a part of AOL/Patch, I’m sure their technology will be put to good use.)

6. Aggregation and syndication are a smoke screen. During a previous round of layoffs, we attempted to automate the posting of content to some of our sites in Morris County. The idea being to aggregate other blogs and syndicate stories from, the sister newspaper site. It was almost immediately apparent that we’d ripped the soul out of the sites, and they quickly became ghost towns. Sure, they had “content” — and thanks to SEO a trickle of traffic — but not a whole lot of living, breathing residents.

7. Display advertising isn’t enough. We simply didn’t see enough local advertising to ever make our sites totally self-sufficient on display dollars. Howard Owens, prepare your flamethrower. I know that he claims to be able to generate upwards of $100-$150k per year on local advertising in The Batavian, and if this is true, my hat is off to him. But most sites are going to need to great creative, and there are plenty of other revenue streams that hyperlocals can employ. Local-local Daily Deals is one idea that has, of course, been discussed. Also promising: The sort of conference that operates; classes like the kind that we co-hosted with the Citizens Campaign; and commercial businesses like Brownstoner’s Brooklyn Flea. In fact, one of the most promising side-businesses just might be news cafes. Which brings us to …

Freehold InJersey "bureau" in Zebu Forno coffee shop

8. Coffee shops are a natural hyperlocal nexus. Many hyperlocal editors work out of Starbucks and neighborhood joints already, but the aforementioned Colleen Curry took the bold step of opening up a bureau in Zebu Forno in Freehold. Alright, it was really just a computer set up with a “The Journalist Is In” sign atop it. But it proved to be a brilliant way to raise awareness of the site and a place to recruit contributors and interview subjects. We didn’t recruit as many folks as we would have liked, but there’s little doubt it raised our profile in town. The Winnipeg Free Press was so inspired by it — and other similar initiatives — that they opened up their own News Cafe this year. Even though some have tried and failed at this model, like Nase Adresa (“Our Address”) in Czech Republic, they never went so far as to actually merge their newsrooms with their cafes.

9. Host the conversation on Facebook. If I could do it all over again, InJersey’s comments, login, and social features would all be rebuilt on Facebook plugins. That’s not to say you should make the drastic move of ditching your website, like Rockville Central. But there’s no need to build your own social network, as we foolishly attempted.

10. Live in the town. File this under “do as we say, not as we do.” For the most part our reporters didn’t live in the towns they covered. Patch, on the other hand, makes it a requirement that you live where you report. Patch’s way is better.

Back when I was prototyping the site in the summer of 2009 with my colleague, Whitney Rhodes (now at Patch), I took great inspiration from Mark Potts’s post-mortem on Backfence (a site whose cemetery I’d be honored to share). His insights guided us in how we approached the site — both from an editorial and community-building standpoint. I hope this will do the same for future hyperlocal entrepreneurs. If we must dub such sites as failures, I just hope we can also see things as Henry Ford did: “Failure is simply the opportunity to begin again, this time more intelligently.”

Read More

Ted’s Twitter Updates for 2011-04-04

Read More