In a surprise move, Xero acquired the Canadian document fetching and processing app HubDoc this week. Xero said it would integrate its fetching technology into the accounting app and continue to operate HubDoc as a standalone product, much as it has done with project management tool WorkflowMax.
It’s an interesting development because of where it puts Receipt Bank, the number one app in the Xero ecosystem. Xero has recently updated Xero Expenses, which shares many functions of the Receipt Bank app, and now it has snapped up one of Receipt Bank’s biggest competitors. (Receipt Bank is beta testing a fetching technology similar to HubDoc’s.)
I spoke with Jamie McDonald, co-founder of HubDoc, on why the acquisition took place and what it means for Xero and HubDoc.
Digital First: What did Xero most want to buy HubDoc? The press release mentioned the document fetching technology right up the top, so I guess that was it?
McDonald: I think the first thing they’re buying is our business, which is a representation of the brand the customer relationships that we have, our team and the growth opportunities there.
The second thing I think is it’s about really overlapping product visions. This idea of HubDoc delivering clean data into the Xero platform to build out a really rich data set. And the idea of delivering an invoice to HubDoc and into Xero, and automatically reconciling bank transactions – I think that is the other big part of it.
Digital First: You mentioned growth opportunities first – is this acquisition about picking up customers in North America to help Xero crack that market?
McDonald: No, I wouldn’t say that. I would say there’s just great momentum behind our business and our brand. Our revenue is growing really fast, our customer base is growing really fast, both in the US, Canada, Australia, and now the UK. So there’s just value in the business and the plan is to continue to operate the brand and run it standalone while we’re considering how to execute on this shared product and a much deeper integration for our customers.
Digital First: How fast are you growing? Can you share customer numbers?
McDonald: We’re not really disclosing any of the numbers as part of the transaction so I’m not going to do that.
Digital First: Ok. Do you think it makes sense to keep the apps standalone or take that fetch technology and add it to Xero?
McDonald: I think it’s an interesting idea [thinking about] what do we actually do at the end of the day? We’re a document ingestion machine or a data extraction machine and then we create a workflow to take that data to automate bookkeeping processes. There’s an interesting debate over where the logical place for that technology lives. Is it integrated deep into the GL over time? Or is there a standalone business? I think the answer might be both.
Digital First: Do you see this acquisition as part of a swing from point applications or best of breed to building a suite under the Xero brand?
McDonald: Yeah, it’s a really good question. I think probably best to ask Steve or anybody else at Xero. I don’t think this is about a roll-up of best of breed apps to build the Microsoft Office for accountants. I think this is one particular horizontal application that’s applicable to every single business on the planet, right?
There’s an opportunity to integrate more deeply between what we do, which is kind of a document and data management platform with the GL. It’s all [about] delivering a better customer experience which manifests itself in how quickly can I get my monthly reconciliation work done and how accurately can I get it done for an accounting firm?
Digital First: So the focus of this acquisition is about automating that data collection process, right?
McDonald: I would divide the process into three stages. Data and document ingestion, or how do you get [your documents]? So we built our business around automation of the fetching process. The second step is data extraction and document classification. What type of document am I looking for and how quickly and accurately can I get the relevant metadata out of that document? And then the third thing is workflow tools that can take that data and turn it into entries in a general ledger and facilitate the automated reconciliation process.
Digital First: Canada and particularly Toronto have a global reputation for AI and machine learning (ML). Do you think the acquisition was also about hiring talent in those areas?
McDonald: I think that’s a yes. Obviously it’s hard to hire ML and AI talent everywhere. We have two particularly excellent universities at training and developing that data science talent. I’m super proud of the strides that our team has made in a short amount of time in terms of leveraging machine learning to drive the sales for customers and actually have it in production. The first task that that team has worked on and successfully delivered against is automated data extraction from manually uploaded documents. So when you take a photo of the receipt in your pocket it figures out date, amount, tax, all of that, and pulling it out automatically.
Then there’s a whole bunch of stuff around our automated fetch that we’ve applied some ML techniques to as well.
Digital First: How have you used machine learning to improve your fetch, specifically?
McDonald: I’m not going to talk about that. It’s kind of proprietary.
Digital First: How do you see this affecting your relationship with Intuit? I just checked the Intuit apps marketplace and HubDoc is still in there.
McDonald: I think the easy analogy is if you look at Intuit’s Tsheets acquisition, Tsheets would still be integrated with Xero and continues to serve their joint customers. It’s certainly our plan to do the same. We have nothing but positive things to say about our long relationship with Intuit and we hope to serve our Quickbooks customers for a long time in the future.
Digital First: Xero mentioned global expansion. Which countries are you going to next?
McDonald: The first thing to do is to deliver excellence against the core Xero customer base, which would be in Australia, Canada, the US, and New Zealand. There’s work to do there. The nature of automated fetching is that it takes some work to launch in each market. You have to build the connections. So that’ll come in good time.
Digital First: Are you in the UK yet?
We’re in the UK already, it is one of the markets that we’re serving.
Digital First: What is on your roadmap for HubDoc? What were you working on prior to the acquisition?
McDonald: I think going back to those three core processes. How can we improve our automated intake? That’s our fetch technology. There’s all sorts of innovations in that platform that we’re making that are super exciting. There’s automated data extraction, which is for the stuff that’s emailed in to us or manually uploaded – how do we get data entered instantly and accurately?
And then the third thing is what are the workflow and publishing tools that we can unlock to drive efficiency for our customers, who are the accountants and bookkeepers around the world that we serve.
So that’s the core roadmap, and the layer beneath that is we’re able to execute properly on that for millions of small businesses globally. We believe it unlocks this unique data asset that other people simply don’t have access to. How can we unlock payments, alternative lending, “always on” audit, and recommendations through the data asset that we’re building. The partnership and marriage with Xero will accelerate that.
Digital First: What is “always on” audit?
McDonald: Well, it’s just the idea. What we do is proving that the entry in the GL happened, right? So 10 years ago, that was done with a piece of paper with a logo on it exchanged between two businesses, which might have been an invoice. We think in the future that every transaction inside of a general ledger will have what we call verified data. That might be an image of a receipt, it might be a link to a blockchain or a decentralised ledger which proves the transaction happened.
Imagine if every transaction in the GL had that proof that it happened. Why would an auditor have to wait until six months after the year’s over to review the books and say that they’re valid? Why couldn’t they say at the end of a quarter, at the end of a month, up to today, that all of these transactions inside the GL are valid and the financials are up to date?
That’s aspirational and a lot of vision and we haven’t at all figured out but that’s certainly stuff that we’ve been thinking about on the product team at HubDoc.
Digital First: Is it true that with every Xero acquisition the founders each get a free Tesla?
McDonald: I was hoping for one of those Minis with the Xero logo on it.
Digital First: It would have to be electric. Petrol is so desktop.
McDonald: Yeah, an electric Mini with the Xero logo on it. I love that idea.
Image credit: Youtube
I decided to get back on the bandwagon with receiptbank given I lost confidence in xero to deliver themselves on this area, so I can automate for myself and clients, just last week. RB use some pretty heavy sales tactics (discounts if sign up by tomorrow etc), the pricing is pretty onerous on small practices imo and the learning curve and setup takes a bit of time, though I can see its improved from 2 years ago and looks like a great product once in place. The next day xero announces the hubdoc purchase – although on my agenda to review, I’d discounted it given RB is the number 1 app…but with the xero acq, it validates it must be worth a look…I signed up and love what hubdoc is doing…in 10 mins I had all my bank statements filed in folders by account in date order, including statements from suppliers I can get from their portals….plus the user layout is structured as an accountant would like it..it becomes a document repository that can be a reference location in one place..can see excellent features for clients that use aggregators etc on the invoicing side too…still to see how good their receipt processing (suspect not as good as RB) but the app is great and cost wise is great value esp with the accountant promo. Very easy to learn and run with. Will help us provide better service to clients which is what we want.
Thanks for the detailed review, Paul. Yes, I think the folder structure with clients in alphabetical order makes a lot of sense for accountants.
I’ve got to say having played with hubdoc, the receipt processing appears less robust /feature rich compared to receiptbank. Receiptbank has come along way in a couple of years, very polished.
Might need a two pronged approach. Good for the market, more players keeps everyone improving.
Interesting summary Sholto. the “Office 365” idea is moot, especially with the background of Steve Vamos. I would suspect that with Hubdoc being kept as a separate brand in the short term, the deal will give Hubdoc massive exposure in new markets; UK and EMEA. Won’t help in newish markets such as ANZ either.
With a globally experienced board in place at Xero, will be no surprise to see more of these acquisitions to follow.
Thanks Dermot, good points