“The next step for Respondly’s growth is figuring out how to get it into as many hands as possible, and for us, partnering with a larger company like Buffer has a tremendous upside for quickly expanding Respondly’s impact,” says Haines, when asked why the team decided to seek an exit at this point, rather than carrying on developing Respondly themselves.
“Buffer are experts in the domain, have a perfectly complementary product, and are well known for their transparency and integrity. They also have 45,000+ customers who already see the value of investing in great social media tools, that they can introduce Respondly to.”
“Respondly has customers who love the product, and depend on it every day. The paying customer base is growing, but not fast enough to continue running Respondly as a stand-alone venture backed business,” he adds. “When we weighed it up, it made rational sense for everyone involved – the customers, the investors, Buffer, and the Respondly team to sell the product to Buffer.”
From Buffer’s point of view, while the call from Haines was serendipitous in terms of meshing with what some of its customers were asking for, the decision to move ahead and make an offer to acquire Respondly boiled down to conducting a cost/benefit analysis of whether it was better for Buffer to build the feature itself or buy Respondly outright and obtain an off-the-shelf product plus a swathe of customers in one fell swoop.
The assessment process involved modeling the buy or built scenarios using some good old-fashioned spreadsheeting, says Widrich — a “very difficult” process that inevitably involved a lot of guesstimates and extrapolation. An exact science this is not (see also: VC).
“We built this spreadsheet and we modeled these two scenarios — one is build, one is buy. And so the build scenario we would say ‘okay, today is December 15, if we were to start building this today’ — which we weren’t even in a position of even starting it at that point… but you need to make some assumptions. So say we were to start building this today, we don’t have the team right now, so we need to start hiring and build this team — or take it off from other teams and start building this first version.
“And so it takes us maybe two months to hire a full time team. And, say, maybe five to eight people. Then we have that team, and after two months they start building it — then we say if we’re fast we can get a first version out say within four months. And so that means six months from now we might have the first version that we can slowly start to introduce to certain customers. Obviously that first version will be very minimal, so it will take up some time to ramp up.
“So then we went further and we said ‘okay, how do we go from the six months, we start with zero MRR — so zero monthly revenue for that product — and then how fast do we think we can build the revenue out, and build the feature set and also hire more people into that team.’ And so we extrapolated that out. So we start with zero and we started thinking okay within the first month we can probably get to $5,000 in MRR, and that already feels very significant because we probably need to get 100 paying customers in the first month or something like that. Which is in itself, really tough. And then we model that out further.
The goal we set ourselves was how long will it take us roughly to get to $100,000 MRR. And the conclusion that we’ve come to in the build model is that would be half way through 2018.
“The goal we set ourselves was how long will it take us roughly to get to $100,000 MRR. And the conclusion that we’ve come to in the build model is that would be half way through 2018. So after we launch it it would probably take us 18 months or six months from now when we launch it it will take us another 18 months roughly to get to $100k MRR. And the way we came to that was we looked at Buffer’s own data that we had — how long it took us to get to that number, and factoring in ‘okay we’ve gotten smarter since then. We know more, we have a customer base we can market to so we can probably do it a lot faster.’ That brings us to this time period of — I think — April 2018,” says Widrich, talking through Buffer’s thought process on the build option.
Another factor it assessed here was associated cost. “Every person that works on that [respond product] team is cost, especially the first six months when we don’t even make any money from that,” he notes. “We’re just basically burning money. Say we have eight people, each cost roughly $120,000/$100,000 a year and so for six months say that’s $60k per person times eight. So that’s like $480k just for the first six months of development.
“The investment, I think we came to, is probably close to I think $1.5M to $2M — and that assumes we will eventually get to $100k MRR… That may never happen. We may never get traction and we just wasted a bunch of money, so we may never get that money back.”
On the buy opportunity side, Buffer’s model considered development cost savings and the risk reduction element given Respondly is an existing product with some customers — off-setting all that against the cost of the acquisition (whatever that is) — so not only the six month dev time it had estimated it would need to make a similar product itself but also some of the initial marketing and traction costs, given that Respondly’s existing customer base would reduce the launch risk as Buffer wouldn’t need to start the business from scratch.
“We said when can we get $100k MRR with the buy model… and so we modeled that out and said ‘okay now is December 15, if we go live with this in January it means six months from now — which we would do in the build model — now we already save six months and we get paying customers and we already have paying customers.’ We also factored in the risk factor — the risk factor goes down, because the likelihood is if you have some paying customers you will keep getting some paying customers is much higher than if you have nothing and you make a bet that you will get at least one paying customer. So the risk from zero to one is much higher than if you already have some.”
Respondly being a relatively mature/feature-rich product was a highly attractive factor for Buffer, adds Widrich. “It would have taken us a long time to add those features. They’ve been working on this for two years, non stop… And to us that was really special because we feel like we could put all the marketing behind it and yes we will keep improving the product of course, but they’d built such an amazing foundation that it’s almost like a ramp that we could just jump off,” he adds.
“And so in this model we assumed that we launch in January and we can get to $100k MRR by the end of 2016. That was our intuition… And we would probably stop losing money half way through 2016 — and we would only burn, I think we figured out roughly, $300k to $400k… So now we’re in a position where we feel like ‘wow, this is actually an amazing place to be in where not only could we save a lot of money — and we wanted to go after this opportunity anyway — but at the same time we are also able to take away a lot of the risk. So there was this compounding effect of that… Therefore the buy option was a lot more attractive for us because of the less investment required from our side plus the lower risk.”
Of course all that thinking is dependent on the value of the acquisition — which again we don’t know. But clearly the size of Buffer’s offer for Respondly undercut its internal calculations on how much it would have cost it to build a similar product itself. (And the associated risks of launching something new — including competing with Respondly (there were other offers on the table for the startup, although Buffer’s offer was apparently the highest.))
So given Widrich’s explanation we can assume the price tag for acquisition was relatively close to the projected cost estimate for building the product itself — aka between $1.5M to $2M. It’s also safe to assume it certainly didn’t exceed Buffer’s own valuation — the latter raised a $3.5 million Series A round itself, back in October 2014, at a $60 million valuation (when it was also profitable and reporting monthly revenue of $400,000 — although Widrich now says it basically breaks even every month as it reinvests in the product).
Buffer’s latest MRR is $663k. It had raised $400,000 prior to the Series A. And had just over $2.5M in cash in the bank as of last month (yes it publishes its bank balance too). Widrich confirms the acquisition was an all-cash deal. So again, we can pretty safely assume the Respondly acquisition price-tag was closer to $2M than $60M given how much cash it has to hand to fund an acquisition.
“[The price of the deal is] obviously the key piece to all of this,” he adds. “It’s something that has to fit — we can’t sort of run ourselves into the ground with an amount that would kill us. For us it became something that an amount that we felt was reasonable and that they were comfortable with. One of the things that we’re always keen to do — because we’re so transparent — is to lay out our way of thinking about this and say look we don’t want to screw anyone over here, we don’t want to try and squeeze out every last dollar from you.
We do some crazy things that people probably advise people never to do in an acquisition.
“The way we approached the interaction with Tim was to be as open and as keen to lay out our way of thinking with him. ‘This is what we feel we can pay’, and we show him our bank balance. We do some crazy things that people probably advise people never to do in an acquisition. But we feel like that’s the best way for us and also to have him build trust in us — that we’re not trying to play games here, but we’re just approaching this in a very rational way. And that helped us to get him on side as well.”
None of the Respondly team will be joining Buffer, so this is a pure-play product acquisition not an acqui-hire. Widrich says it’s assigning nine people to work on the Respond product as it takes it over from Respondly. It’s ramped up seriously on the staff front this year, growing from around 25 employees at the start of the year to almost 70 now.
So how did Buffer assess the risk of integrating an existing product into its existing business — and doing so without any of Respondly staff on board to smooth the transition? Calculating that unknown inevitably involved a lot more conjecture.
“We’ve taken a growth rate and we’ve taken certain numbers we think this new product can grow into and we have then applied a risk factor — sort of like decreased the growth rate based off a certain risk factor we think is existing,” says Widrich. “To give you one example, one risk is say for example Twitter. Twitter’s API and their approach to developers. What if we buy Respondly and then next day we get cut off from their API so we can’t even provide the product anymore?
“That’s something we should factor in. And some of those big ticket items are really hard to calculate but the best way we’ve found to do this is to apply a percentage and say ‘hey actually let’s decrease all of this by 25 per cent to account for unforeseen things that may happen that throw us off and mean that we need more time.’ That’s the best way we’ve been able to do this.
“And if there are very specific risk factors that we can already identify — obviously we can’t look into the future — for example that we’re dependent on a platform like Twitter or that.. say something else changes in the market. We can think of Facebook working on their messenger tool, turning that into customer service — and that can be a good thing or that can hurt us. So we’re trying to list some of those things that apply and say let’s be more conservative about our forecast. So that we have some slack in terms of the numbers that we want to hit or we can get to.”