\n",
"\n",
"Once [again](http://www.columbia.edu/~srh2144/fun.html#balls) my answer is too complicated. Regularized logistic regression is a similar model which gives pretty much the same answer. [Update discussion.](http://nbviewer.ipython.org/github/sharnett/AB-testing/blob/master/Updated%20thoughts%20on%20Bayesian%20AB%20Testing.ipynb)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Intro\n",
"\n",
"The typical approach to Bayesian AB testing doesn't really allow for prior knowledge on the scale of the change in conversion rate. Sometimes you know that a change will probably have only a small effect. I propose a slightly different approach that explicitly models knowledge about the expected effect size."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#The usual way of thinking about it\n",
"\n",
"We have two different versions of a website, A and B, and we wish to estimate their true conversion rates $p_A$ and $p_B$. (Or more likely we want an estimate of the difference or ratio of the conversion rates, with some notion of the uncertainty around that estimate.) We observe a number of visits, some of which convert, to each version of the site. Call these $obs_A$ and $obs_B$. We also have some prior knowledge of the possible values for $p_A$ and $p_B$ which we model as beta distributions $B_A$ and $B_B$. A typical choice for these is the flat, uniform prior for each: $B(1,1)$.\n",
"\n",
"So the data story goes like this for each variation:\n",
"\n",
"1. Select the conversion rate $p$, where $p$ is beta distributed according to the prior $B$.\n",
"2. Observe a number of visits which are Bernoulli distributed with rate $p$.\n",
"\n",
"$$B \\rightarrow p \\rightarrow obs$$"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#A problem with this story\n",
"\n",
"Often the differences between A and B are small, e.g. rephrasing the call to action on a button. So you have prior knowledge that the difference in conversion rates $p_A$ and $p_B$ is probably small. The above story doesn't really allow you to use that knowledge. The problem is worse the flatter your priors $B_A$ and $B_B$ are. In the case of a uniform prior for both, we're saying that $p_A, p_B = 0\\%, 100\\%$ is equally likely to $p_A, p_B = 25\\%, 25\\%$, which is crazy for a small change."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#A better data generation story\n",
"\n",
"I use an [alternative parameterization](http://en.wikipedia.org/wiki/Beta_distribution#Alternative_parametrizations) of the beta distribution, based on the mean $\\mu = \\frac{\\alpha}{\\alpha+\\beta}$ and \"sample size\" or spread $\\tau = \\alpha + \\beta$.\n",
"\n",
"Rather than using a fixed beta prior for A and B, we allow it to vary according to our prior knowledge. The mean $\\mu$ varies according to historical data, much like in the usual approach. Denote the varying beta prior $B_{\\mu}$ to indicate it's dependence on $\\mu$. Fix the spread $\\tau$ to a value which is consistent with our beliefs about the possible effect size. $p_A$ and $p_B$ are then drawn from this beta distribution.\n",
"\n",
"The story looks like this:\n",
"\n",
"1. Based on historical data, select a beta distribution $B_0$\n",
"2. Draw a $\\mu$ from $B_0$\n",
"3. Based on intuition/experience, choose a spread parameter $\\tau$\n",
"4. We now have a beta distribution $B_{\\mu}$\n",
"5. Select $p_A$ and $p_B$ from $B_{\\mu}$\n",
"6. Observe visits that are Bernoulli distributed according to $p_A$ and $p_B$\n",
"\n",
"In picture form:\n",
"\n",
"![Figure 1](files/data_story.png \"Figure 1\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#Tuning $\\tau$\n",
"\n",
"The step for coming up with $\\tau$ sounds a bit magical in the above description, so let's be more concrete. Suppose that from our historical data, the mean conversion rate for situations similar to our current experiment is 10%, so $\\mu=.1$. Setting $\\tau=1500$ then produces a distribution of the conversion rate difference shown below. Increasing $\\tau$ narrows the distribution. Use a bit of guess and check to come up with something reasonable here."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"import numpy as np\n",
"from scipy.stats import beta as beta_dist\n",
"%matplotlib inline"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 1
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"mu = .10\n",
"tau = 1500\n",
"alpha, beta = mu*tau, (1-mu)*tau\n",
"pA = beta_dist.rvs(alpha,beta,size=10000)\n",
"pB = beta_dist.rvs(alpha,beta,size=10000)\n",
"diff = Series(pA-pB)\n",
"diff.hist(bins=np.arange(-.05,.06,.01), histtype='stepfilled')\n",
"print(diff.describe())"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"count 10000.000000\n",
"mean 0.000137\n",
"std 0.010988\n",
"min -0.042673\n",
"25% -0.007296\n",
"50% 0.000271\n",
"75% 0.007532\n",
"max 0.037420\n",
"dtype: float64\n"
]
},
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAECCAYAAAARlssoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKIiIiKFQEBERQ6EgIiLG/wfMR22zc8Yz2wAAAABJRU5ErkJggg==\n",
"text": [
""
]
}
],
"prompt_number": 2
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#Example\n",
"\n",
"In the example below, it's early in the AB test and we only have a few thousand sessions. But the results look very promising, almost a 15% lift in the empirical conversion rate. But how realistic is that, if it was only a small change to the website? Should we stop the test and call it a win?\n",
"\n",
"I take a look at the estimated percentage improvement in conversion rate with the usual method versus the new method. For the usual method, I try it with a flat prior $B(1,1)$ as well as the very informative prior $B(30, 270)$.\n",
"\n",
"The new method gives a much more modest estimate of the improvement."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"successesA = 300\n",
"successesB = 350\n",
"failuresA = 3000\n",
"failuresB = 3000\n",
"pA, pB = successesA/(successesA+failuresA), successesB/(successesB+failuresB)\n",
"pA, pB, (pB-pA)/pA"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 3,
"text": [
"(0.09090909090909091, 0.1044776119402985, 0.1492537313432835)"
]
}
],
"prompt_number": 3
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def usual_method(alpha=1, beta=1):\n",
" pA = beta_dist.rvs(alpha+successesA,beta+failuresA,size=10000)\n",
" pB = beta_dist.rvs(alpha+successesB,beta+failuresB,size=10000)\n",
" lift = Series((pB-pA)/pA)\n",
" return lift"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 4
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def new_method(alpha0=3, beta0=27, tau=1500):\n",
" mu_samples = beta_dist.rvs(alpha0, beta0, size=10000)\n",
" def compute_p(mu):\n",
" alpha, beta = mu*tau, (1-mu)*tau\n",
" pA = beta_dist.rvs(alpha+successesA,beta+failuresA)\n",
" pB = beta_dist.rvs(alpha+successesB,beta+failuresB)\n",
" return pA, pB\n",
" p = np.array([compute_p(mu) for mu in mu_samples])\n",
" pA, pB = p[:, 0], p[:, 1]\n",
" lift = Series((pB-pA)/pA)\n",
" return lift"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 5
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"lift0 = usual_method()\n",
"lift1 = usual_method(30, 270)\n",
"lift2 = new_method()\n",
"df = DataFrame({'Flat prior': lift0, 'Very informative prior': lift1, 'New method': lift2})\n",
"df = pd.melt(df)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 6
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"If we're using a simple decision rule like \"stop the test if one variation has a 95% or greater of being the best\" then the usual method would stop here, but the new method would want to gather more data:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"(lift0<0).mean(), (lift1<0).mean(), (lift2<0).mean()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 7,
"text": [
"(0.031800000000000002, 0.035200000000000002, 0.059999999999999998)"
]
}
],
"prompt_number": 7
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Even using an unrealistically informative prior doesn't shift our estimated CVR lift too much. But the new method does."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"%load_ext rmagic"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 8
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"%%R -i df -w 900\n",
"library(ggplot2)\n",
"p <- ggplot(df, aes(x=value, color=variable)) + geom_density()\n",
"p + xlab('Estimated conversion rate lift') + xlim(-.1, .4)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAA4QAAAHgCAYAAAD9tKNeAAAEJGlDQ1BJQ0MgUHJvZmlsZQAAOBGF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\n"
}
],
"prompt_number": 9
}
],
"metadata": {}
}
]
}